JP6634938B2 - Analysis support method, analysis support program, and analysis support device - Google Patents

Description

  An embodiment of the present invention relates to an analysis support method, an analysis support program, and an analysis support device.

  2. Description of the Related Art Conventionally, in a business system using a database (hereinafter, referred to as a DB) such as a relational database (RDB), when maintenance and development such as addition of a new function is performed, the influence of a change due to the addition of a new function spreads to what extent. It is required to know what. In grasping the range affected by the influence, it is investigated how data is propagated by writing and reading data to and from the DB. For example, if the DB is an RDB, writing corresponds to an INSERT or UPDATE instruction in SQL, and reading corresponds to a SELECT instruction.

  Methods for investigating how data is propagated by writing to the DB with an INSERT or UPDATE instruction and reading with a SELECT instruction are roughly classified into a static method and a dynamic method.

  As a static method, there is a method of analyzing a flow of data when a business system is executed from a source code related to the business system. By using the source code, data propagation in the source code can be accurately grasped. However, when there is input / output of external data such as a DB in a business system, it is necessary to find another source code related via the DB and to grasp data propagation in the source code.

  For example, when SQL is issued from a method described in source code such as Java (registered trademark) of a business system to exchange data with an RDB, both SQL and source code are analyzed in order to grasp data propagation. There is a need to. However, correctly analyzing the source code is troublesome, and it is difficult to grasp the data propagation. Also, in the case of the static method, it is difficult to grasp the order relation in which writing or reading to certain data (column) occurs in actual work, and it is possible to extract data propagation that does not actually occur. There is.

  On the other hand, as a dynamic method, there are a method of acquiring and analyzing an operation log related to program execution, and an analysis using an SQL execution history as an operation log. The operation log to be analyzed includes, for example, the executed SQL and the parameters at that time, and in the case of a SELECT command, a search result.

JP 2010-102664 A

  However, the above-described related art has a problem that it is sometimes difficult to accurately grasp data propagation in a business system using a DB. For example, it is difficult to accurately grasp indirectly related data propagation in which a value read from a DB is operated in a method and the value after the operation is written in another column.

  FIG. 27 is an explanatory diagram illustrating the business system 300. As shown in FIG. 27, the business system 300 includes a Web server 311, a business server 312, a business DB 313, and DB access data 314.

  The Web server 311 receives an access from the browser of the terminal device 301 and provides a business process service to the browser of the terminal device 301. Specifically, the Web server 311 notifies the business server 312 of a processing request via the HTTP (HyperText Transfer Protocol) parameter from the terminal device 301, and returns the processing result of the business server 312 to the browser of the terminal device 301. The business server 312 executes a business process by accessing a business DB 313 which is an RDB storing various business information based on a processing request from the Web server 311 and executes a business process, and transmits a processing result to the Web server 311. return. Further, the business server 312 stores a log (operation log) relating to access to the business DB 313 in the DB access data 314. The operation log includes the executed SQL and parameters, and a search result in the case of a SELECT command.

  For example, it is assumed that the access from the Web server 311 triggers the business server 312 to execute the SELECT statement and then execute the INSERT statement. In this case, based on the operation log of the DB access data 314, the HTTP parameter (screen parameter), the SELECT statement and the parameter, the search result, the parameter of the INSERT statement, the execution time, Is obtained.

  In this operation log, for example, the value (y1) read out from the column C11 of Table1 in the business DB 313 by the SELECT statement and the value (z1) written in the INSERT statement to the column C21 of Table2 in the business DB 313 are always stored in a dynamic manner. It is assumed that y1 = z1 is satisfied. In this case, it can be estimated that the value of column C11 has propagated to column C21 via the business process. As described above, when the value of the search result is directly written into another column by the INSERT statement, data propagation can be grasped by a dynamic method.

  However, when the read value (y1) is calculated in the method and the calculated value (z1) is written to another column, since y1 ≠ z1, an indirect relationship exists. Nevertheless, it is difficult to accurately grasp the relationship. In the case of updating with an UPDATE statement, for example, a SELECT statement is executed once, an operation is performed on the read value, and the value after the operation is updated with the UPDATE statement. As described above, even when the update is performed by the UPDATE statement, since the existing value is not used as the update value as it is, even though there is an indirect relationship, the relationship is accurately grasped. It is difficult to do.

  In one aspect, an object of the present invention is to provide an analysis support method, an analysis support program, and an analysis support device capable of accurately grasping data propagation in a business system.

  In the first proposal, the analysis support method includes a computer which includes an input parameter, and an access record in which at least one of reading data from the database and writing data to the database is written based on the input parameter. Execute the process of acquiring a plurality of operation logs including In the analysis support method, the computer may be configured such that in each of the plurality of acquired operation logs, in each of the operation logs including the access record related to a predetermined column of the database, the value included in the input parameter or the database If the calculation result obtained by performing a predetermined calculation in combination with the value read from the data and the value written to the database in the access record is constant, it is analyzed that there is data propagation in the column. Also, in the analysis support method, the computer executes a process of outputting a column analyzed when there is propagation.

  According to one embodiment of the present invention, data propagation in a business system can be accurately grasped.

FIG. 1 is a block diagram illustrating the configuration of the analysis support apparatus according to the embodiment. FIG. 2 is a flowchart illustrating an operation example of the analysis support device according to the embodiment. FIG. 3 is a flowchart showing the processing of “association with one column”. FIG. 4 is a flowchart showing the processing of “association with one screen parameter”. FIG. 5 is a flowchart showing the process of “estimating the relationship between a plurality of repetition parameters of the screen parameter and the INSERT value”. FIG. 6 is a flowchart showing the process of “estimating the relationship between the two columns of the SELECT result and the INSERT value”. FIG. 7 is a flowchart showing the process of “estimating the relationship between two types of SELECT result columns and INSERT values related via key values”. FIG. 8 is a flowchart illustrating the process of “estimating the relationship between the screen parameter associated with the key value, the column of the SELECT result, and the INSERT value”. FIG. 9 is a flowchart illustrating an operation example of the analysis support device according to the embodiment. FIG. 10 is a diagram illustrating a table configuration of the business DB. FIG. 11 is an explanatory diagram illustrating an example of an operation log. FIG. 12 is an explanatory diagram illustrating an example of the data table. FIG. 13 is an explanatory diagram illustrating an example of the operation log. FIG. 14 is an explanatory diagram illustrating an example of a data table. FIG. 15 is an explanatory diagram illustrating an example of an operation log. FIG. 16 is an explanatory diagram illustrating an example of a data table. FIG. 17 is an explanatory diagram illustrating an example of an operation log. FIG. 18 is an explanatory diagram illustrating an example of a data table. FIG. 19 is an explanatory diagram illustrating an example of the operation log. FIG. 20 is an explanatory diagram illustrating an example of the data table. FIG. 21 is an explanatory diagram illustrating an example of the operation log. FIG. 22 is an explanatory diagram illustrating an example of the data table. FIG. 23 is an explanatory diagram illustrating an example of an operation log. FIG. 24 is an explanatory diagram illustrating an example of a data table. FIG. 25 is an explanatory diagram illustrating an example of a display screen. FIG. 26 is an explanatory diagram illustrating an example of a computer that executes an analysis support program. FIG. 27 is an explanatory diagram illustrating the business system. FIG. 28 is an explanatory diagram illustrating an example of data propagation in an order receiving business. FIG. 29 is an explanatory diagram illustrating a case where detection of data propagation has failed. FIG. 30 is a flowchart illustrating an operation example of the analysis support device according to another embodiment. FIG. 31 is an explanatory diagram illustrating an example of an operation log. FIG. 32 is an explanatory diagram illustrating an example of a data table. FIG. 33 is an explanatory diagram illustrating an example of a data table. FIG. 34 is an explanatory diagram illustrating an example of a data table. FIG. 35 is an explanatory diagram illustrating detection of data propagation. FIG. 36 is an explanatory diagram illustrating an example of a data table. FIG. 37 is an explanatory diagram illustrating an example of a data table. FIG. 38 is an explanatory diagram illustrating an example of the data table. FIG. 39 is an explanatory diagram illustrating detection of data propagation.

  Hereinafter, an analysis support method, an analysis support program, and an analysis support device according to an embodiment will be described with reference to the drawings. In the embodiments, configurations having the same functions are denoted by the same reference numerals, and redundant description will be omitted. The analysis support method, the analysis support program, and the analysis support device described in the following embodiments are merely examples, and do not limit the embodiments. In addition, the following embodiments may be appropriately combined within a range not inconsistent.

  FIG. 1 is a block diagram illustrating the configuration of the analysis support device 1 according to the embodiment. As shown in FIG. 1, the analysis support apparatus 1 is an information processing apparatus such as a PC (Personal Computer) or a workstation, acquires an operation log of the business system 2 stored in the DB access data 20, and acquires the acquired operation log. Perform log analysis. The analysis support device 1 includes an operation log acquisition unit 10, an analysis unit 11, a storage unit 12, and an output unit 13.

  In the business system 2, an operation log obtained when the business server accesses the business DB and executes the business process in response to a request from the terminal device to the Web server that provides the business processing service is used as a log file as the DB access data 20. For example. In this embodiment, an analysis of an operation log of online execution in response to a request from a terminal device via a Web server will be exemplified. Specifically, the operation log describes the history of the business server accessing the business DB and executing the business process using the value input from the screen on the browser. Note that the content of the operation log to be analyzed is not limited to the above example. For example, the present invention can also be applied to the analysis of an operation log when a business server accesses a business DB by batch processing and executes business processing.

  Further, in the present embodiment, it is assumed that the SQL statement relating to the access to the business DB is described using a bind variable in the DB access data 20. With respect to this SQL statement, it is possible to extract a portion corresponding to a bind variable from a variable value embedded therein and convert it into an SQL, a bind variable, and a bind variable value.

  The operation log acquisition unit 10 acquires an operation log of the business system 2 by accessing a predetermined log file in the business system 2 to be analyzed. Specifically, the operation log acquisition unit 10 inputs, as an operation log of the business system 2, an HTTP parameter (screen parameter) input on the screen of the terminal device and input to the business server, and a predetermined value of the business DB based on the input. And the information (processing time, etc.) that can determine the access content to the column and the execution order of these. The contents of access to a predetermined column of the business DB include a series of SQL logs (SQL statements, input parameters (bind variable values), and SELECT statements in the case where the screen parameters are input from the screen of the terminal device). Is in addition to the search results).

  In the following, a set of operation logs (screen parameters and SQL logs) in a business process that has been triggered by input from one screen of the terminal device will be referred to as a screen unit log below.

  Based on the contents of the operation log acquired by the operation log acquisition unit 10, the analysis unit 11 transfers the variable values in the screen parameters and the values read from a certain column of the business DB to a predetermined column of the business DB. Analyze whether there is data propagation to propagate. The storage unit 12 stores work data (such as the operation log acquired by the operation log acquisition unit 10) when performing the process analyzed by the analysis unit 11. The output unit 13 outputs the analysis result of the analysis unit 11 on a display or printed by a printer.

  The analysis unit 11 inputs an input variable value such as a variable value in a screen parameter or a value read from a certain column of the business DB based on an operation log for a predetermined column, and an output to be written to the predetermined column. It is determined whether or not the variable value matches, and whether or not there is direct data propagation to a predetermined column is analyzed. In addition, the analysis unit 11 calculates a value read from the business DB based on a plurality of operation logs for a predetermined column, and writes the calculated value to another column. Analyze for presence. Specifically, in a plurality of operation logs for a predetermined column, the analysis unit 11 performs a predetermined operation when a result of performing a predetermined operation by combining an input variable value and an output variable value is constant. Analyze that there is indirect data transmission to the column.

  As for indirect data propagation, it is difficult to analyze the presence or absence of data propagation for an arbitrary operation. Therefore, in the present embodiment, a predetermined calculation that is likely to occur in business processing is targeted. More specifically, it targets a constant multiplication operation (eg, calculation of consumption tax), a product-sum operation (eg, calculation of total amount from unit price and quantity), and addition of a fixed value.

  Here, the analysis of the presence or absence of indirect data propagation will be described in detail. First, when writing to a predetermined column, there are a case where a value is inserted with an SQL INSERT statement and a case where a value is updated with a SQL UPDATE statement.

(In the case of INSERT)
The analysis unit 11 extracts, for each screen unit log, one or more identical SELECT statements (but different parameters) executed before the INSERT statement. When the same SELECT statement exists, the analysis unit 11 calculates the next value from the search result or one or two screen parameters.
・ One column or screen parameter: α ・ C + β
・ Two columns or screen parameters: αΣCDD + β
Here, α and β are constants, and C and D are values of certain columns or screen parameters of the table obtained by the SELECT statement, respectively.

  If the calculated value always matches the value to be inserted in the INSERT statement, the analysis unit 11 analyzes that the search result column or screen parameter propagates to the INSERT target column. The calculation of the constants α and β is performed as follows.

(When there is only one column or screen parameter)
The analysis unit 11 calculates the value of a specific column C (here, C1 and C2) of a search result obtained by a certain SELSET from two times of a certain screen unit log (however, screen parameters are different), and The value (here, X1, X2) inserted into the column (X) is obtained by the INSERT executed later.

  Next, the analysis unit 11 calculates a value of α = (X1−X2) / (C1−C2), and determines whether or not the calculated value is always constant regardless of the combination of the operation logs. Calculate using When the value of α is always constant, there is a relationship of β = X1−αC1. Therefore, when the value of α is always constant in a plurality of operation logs, the analysis unit 11 transforms the value of the column C obtained by SELECT into αC + β and inserts it into the column X and propagates it. And analyze.

(Estimation of relationship between one screen parameter and INSERT value)
The analysis unit 11 inserts the values of the specific parameters C (here, C1 and C2) among the screen parameters and inserts them into the column (X) by INSERT from two screen unit logs (however, the screen parameters are different). (Here, X1 and X2) are obtained.

  Next, the analysis unit 11 calculates a value of α = (X1−X2) / (C1−C2), and determines whether or not the calculated value is always constant regardless of the combination of the operation logs. Calculate using When the value of α is always constant, there is a relationship of β = X1−αC1. Accordingly, when the value of α is always constant in the plurality of operation logs, the analysis unit 11 analyzes that the value of the screen parameter C is transformed (calculated) by αC + β, and INSERTed into the column X and propagated. .

(When there are two columns or screen parameters)
Even when there are two columns or screen parameters, the analysis unit 11 uses two screen unit logs to determine the relationship between SELECT results or screen parameters executed before the INSERT to be investigated in the following four cases. analyse.

(Estimation of relationship between multiple repetition parameters of screen parameter and INSERT value)
When the same type of parameter value is repeated a plurality of times in the screen parameter, two types (C and D) of the repeated parameter obtained from a certain operation log and the value (X1) inserted into the column X by INSERT, It is assumed that a parameter value obtained from another operation log on the same screen is obtained. The two kinds of values of the repeated parameter are C1i and D1i (i = 1..m, where m is the number of repetitions of the parameter). The parameter values obtained from different operation logs of the same screen are C2i and D2i (i = 1..n, where n is the number of repetitions of the parameter).

  At this time, the analysis unit 11 calculates whether the value of α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) is always constant irrespective of the combination of the operation logs by using a combination of a plurality of operation logs. When α is constant, the analysis unit 11 analyzes that β = X1−αΣC1i · D1i, and that the values of the screen parameters C and D are transformed into αXC · D + β and inserted into the column X and propagated.

  Note that the parameter value is repeated a plurality of times because, for example, a character string such as “abc- *, def- *, ghi- *,. It is the one that is easily detected as being repeated. As an example, if the URL is "http: //aaa.bbb/exec? Xyz = 001 & abc-1 = 123 & def-1 = 234 & ghi-1 = 345 & abc-2 = 111 & def-2 = 222 & ghi-2 = 333 & zzz = 999", the parameter name If the last number is ignored, the same string appears repeatedly in the same order. In the illustrated URL, the parameter values corresponding to “abc- *” are (123, 111) in order, and the parameter values corresponding to “def- *” are (234, 222) in order.

(Estimation of the relationship between the two columns of the SELECT result and the INSERT value)
In the screen unit log, SELECT (when one or more search results are obtained in one SELECT, or when one or more identical SELECTs with different parameters are executed and one or more search results are obtained) It is assumed that INSERT is executed later. Also, the values of the two columns C and D of the search result (C1i and D1i, respectively, i = 1..m, where m is the number of search results) and the value (X1) set in column X by INSERT ) Is obtained, and the values of C, D, and X (C2j, D2j, j = 1..n, where n is the number of search results, X2) are obtained in another operation log.

  At this time, the analysis unit 11 calculates whether the value of α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) is always constant irrespective of the combination of the operation logs by using a combination of a plurality of operation logs. The analysis unit 11 analyzes that when α is always constant, β = X1−αΣC1i · D1i, and the values of the columns C and D are subjected to the deformation αΣC · D + β and are INSERTed into the column X and analyzed. I do.

(Estimation of the relationship between the two types of SELECT result columns and INSERT values that are related via key values)
In the screen unit log, values of specific columns C and D (C1i and D1i, i = 1, respectively) obtained from two different SELECT search results (however, a combination having the same number of search results are selected) are obtained. M, where m is the number of search results), and the value (X1) inserted into the column X by the INSERT executed thereafter. Similarly, it is assumed that the same column C, D, and X values (C2j, D2j, j = 1... N, where n is the number of search results, X2) are obtained from another operation log. .

  At this time, the analysis unit 11 calculates, using a plurality of operation logs, whether the value of α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) is always constant regardless of the combination of the operation logs. However, a column serving as a key for associating records including the columns C1i and D1i is present in each record, and C1i and D1i are obtained from the record associated with the key value. The same applies to C2j and D2j. When α is always constant, the analysis unit 11 analyzes that β = X1−αΣC1i · D1i, and that the values of the columns C and D have been transformed to αΣC · D + β and propagated to the column X.

(Estimation of the relationship between the screen parameter associated with the key value, the column of the SELECT result, and the INSERT value)
In the screen unit log, when two or more types of the same parameter value are repeated in the screen parameter, there is a SELECT in which the number of repetitions and the number of search results are the same, and one of the screen parameters and the column of the search result have the same key relationship. It is assumed that Also, a specific column C of the SELECT result, a screen parameter D that is not a key (C1i and D1i, i = 1..m, where m is the number of search results), and a column X in the INSERT executed thereafter And the value (X1) inserted in. Similarly, it is assumed that the same values of columns C, D, and X (C2j, D2j, j = 1..n, where n is the number of search results, X2) are obtained from another operation log.

  At this time, the analysis unit 11 calculates, using a plurality of operation logs, whether the value of α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) is always constant regardless of the combination of the operation logs. When α is always constant, the analysis unit 11 determines that β = X1−αΣC1i · D1i, and that the values of the screen parameters C and the SELECT result D are transformed into αΣC · D + β and INSERTed into the column X and propagated. presume.

(In the case of UPDATE)
When setting the value by UPDATE, the analysis unit 11 analyzes data propagation in the following processing. It is assumed that one or more identical SELECT statements (but different parameters) executed before UPDATE exist in the screen unit log. At this time, when the next value calculated from one or two of the search result or the screen parameter always matches the value to be updated in the UPDATE, the analysis unit 11 transmits the column or the screen parameter to the UPDATE target column. Analyze that you did.
・ One column or screen parameter: α ・ C + β
・ Two columns or screen parameters: αΣCDD + β
However, the calculation method of α and β is the same as described above.

  Also, when there is one or more identical SELECT statements (but different parameters) executed before the UPDATE, the next value calculated from the search result or one or two of the screen parameters is updated in the UPDATE. If the increment value (the difference between the newly set value and the immediately preceding SELECT result) always matches, the analysis is performed that the column or the screen parameter propagates to the UPDATE target column.

  In the above description, it is assumed that the SQL string included in the screen log when a certain screen is triggered is always the same. However, the SQL column to be executed may be different depending on the value input from the screen. In this case, the above-described processing is performed by regarding the execution in which the executed SQL group is the same as the same screen.

  FIG. 2 is a flowchart illustrating an operation example of the analysis support apparatus 1 according to the embodiment, and is an example of processing in the case of the above-described INSERT. As shown in FIG. 2, when the process is started, the operation log acquisition unit 10 acquires an operation log from the business system 2 to be analyzed with reference to the DB access data 20 (S1). The acquired operation log includes a screen parameter (HTTP parameter), an SQL statement, an SQL parameter, a search result when the SQL is SELECT, and a time at which each was executed (the accuracy with which the execution order can be determined). The operation log acquiring unit 10 extracts a query character string (a pair of a parameter name and a value) described in a URL or a message body, which is transmitted by a GET or POST method request, from the HTTP parameter.

  Next, the analysis unit 11 divides the operation log acquired by the operation log acquisition unit 10 into screen unit logs on a screen basis based on the screen parameters. For example, the analysis unit 11 divides the operation log for each access to a URL that displays the same screen. Then, the analysis unit 11 extracts an SQL sequence from the divided screen unit log (S2).

  Next, the analysis unit 11 creates a set of screen-based operation logs having the same combination of executed SQL statements based on the SQL string extracted for each screen-based log (S3). At this time, it is assumed that the difference in the number of executions of the same SQL statement (SQL parameters are different) is not considered.

  Next, the analysis unit 11 executes the first loop processing of S4 to S14 for each set of the created operation logs for each screen. When the first loop process is started, the analysis unit 11 associates the columns of the table to be INSERTed with the columns obtained by the SELECT result executed before, using a known method (S5). Specifically, the analysis unit 11 has a case where the value (y1) read from a certain column by the SELECT statement and the value (z1) written by another in the INSERT statement to another column always satisfy y1 = z1. Analyze that there is direct data propagation from one column to another.

  Next, the analysis unit 11 executes the second loop processing of S6 to S13 on the column (X) that has not been analyzed in S5 as having direct data propagation and has not been related.

  In the second loop process, the analysis unit 11 executes a process of “association with one column” (S7) and a process of “association with one screen parameter” (S8). In addition, the analysis unit 11 executes a process of “estimating a relationship between a plurality of repetition parameters of the screen parameter and the INSERT value” (S9) and a process of “estimating a relationship between two columns of the SELECT result and the INSERT value” (S10). I do. In addition, the analysis unit 11 performs the process of “estimating the relationship between the two types of SELECT result columns and the INSERT value that are related via the key value” (S11), and the “screen parameter and the SELECT result column that are related via the key value. (S12) of "estimation of relationship between INSERT and INSERT value".

  When the first and second loop processes are completed, the output unit 13 outputs the analysis result of the analysis unit 11 by the first and second loop processes on a display or by printing by a printer (S15). Specifically, based on the analysis result of the analysis unit 11, the output unit 13 determines a predetermined column of the business DB and a variable value in a screen parameter analyzed as having data to be transmitted to this column. An image indicating the correspondence with the column of the business DB is output.

  As an example, the output unit 13 displays an image in which columns and variable values having a corresponding relationship are connected by a line in a table listing the columns of the business DB and the variable values in the screen parameters. The output mode of the image showing the correspondence between the columns and variable values analyzed with direct data propagation, and the correspondence between the analyzed columns and variable values with indirect data propagation. May be different from each other. For example, when the output unit 13 displays an image in which columns and variable values having a corresponding relationship are connected by a line, columns and variable values analyzed as having direct data propagation are displayed as solid lines, Columns and variable values analyzed as having indirect data propagation may be indicated by broken lines. In this way, by making the output form of the analysis result analyzed as having direct data propagation different from that of the analysis result having indirect data propagation different from each other, the analysis result is obtained. It can be easier to see.

  FIG. 3 is a flowchart showing the processing of "association with one column" (S7). As shown in FIG. 3, when the processing of S7 is started, the analysis unit 11 executes a SELECT statement executed from the set of operation log to be processed until the screen is triggered to INSERT into the column (X) to be processed. A set of columns (SetC) included in the search result is created (S20).

  Next, the analysis unit 11 executes a loop process of S21 to S26 for each column (C) included in the column set (SetC). When the loop process is started, the analysis unit 11 selects one screen unit log and acquires the value (c1Val) of the column (C) of the SELECT result and the value (x1Val) of the column (X) to be inserted by INSERT. (S22).

Next, the analyzing unit 11 uses the value (c2Val) of the column (C) of the SELECT result and the value (x2Val) of the column (X) to be inserted by INSERT for each screen unit log other than the log used in S22. Is calculated to determine the value of α (S23).
α = (x1Val−x2Val) / (c1Val−c2Val)

  Next, the analysis unit 11 determines whether or not α is the same in all the operation logs (screen unit logs) (S24). If α is the same value (S24: YES), the analysis unit 11 analyzes that the value of the column (C) has been subjected to a predetermined operation, and has been inserted into the column (X) and propagated. Therefore, the analysis unit 11 estimates that there has been indirect data propagation from the column (C) to the column (X) (S25), and loops the processing to the next column (S26). If α is not the same value (S24: NO), the analysis unit 11 does not presume that indirect data has propagated from the column (C) to the column (X), and proceeds to the next column. Loop.

  FIG. 4 is a flowchart showing the process (S8) of "association with one screen parameter". As shown in FIG. 4, when the process of S8 is started, the analysis unit 11 creates a set of screen parameters (SetC) from the set of operation logs to be processed (S30).

  Next, the analysis unit 11 performs a loop process of S31 to S36 for each screen parameter (C) included in the created set (SetC). When the loop process is started, the analysis unit 11 selects one screen unit log and acquires the value (c1Val) of the screen parameter (C) and the value (x1Val) of the column (X) to be inserted by INSERT (S32). ).

Next, the analysis unit 11 calculates the next calculation using the value (c2Val) of the screen parameter (C) and the value (x2Val) of the column (X) to be inserted by INSERT for each screen unit log other than the log used in S32. To obtain the value of α (S33).
α = (x1Val−x2Val) / (c1Val−c2Val)

  Next, the analysis unit 11 determines whether or not α is the same in all operation logs (screen unit logs) (S34). When α is the same value (S34: YES), the analysis unit 11 analyzes that the value of the screen parameter (C) has been subjected to a predetermined operation and has been inserted into the column (X) and propagated. Therefore, the analysis unit 11 estimates that there has been indirect data propagation from the screen parameter (C) to the column (X) (S35), and loops the processing to the next screen parameter. If α is not the same value (S34: NO), the analysis unit 11 does not estimate that there has been indirect data propagation from the screen parameter (C) to the column (X), and proceeds to the next screen parameter. And loop the process.

  FIG. 5 is a flowchart illustrating the process (S9) of “estimating the relationship between a plurality of repetition parameters of the screen parameter and the INSERT value”. As shown in FIG. 5, when the process of S9 is started, the analysis unit 11 acquires a set of parameter names that appear the same number of times from the screen parameters (S40). Next, the analysis unit 11 creates a combination of two parameters for each pair of parameter names that appear the same number of times (S41).

  Next, the analysis unit 11 executes a loop process of S42 to S47 for each created combination. When the loop processing is started, the analysis unit 11 selects one screen unit log, and calculates the values of two parameter names (C1i and D1i, i = 1..m, respectively, where m is the number of repetitions of the parameter). Take out. Then, the analysis unit 11 calculates the sum of products (ΣC1i · D1i) (S43). Further, the analysis unit 11 acquires the value (X1) inserted into the column (X) by INSERT.

Next, for each screen unit log other than the one used in S43, the analysis unit 11 determines, from the screen parameters, the values of each of the two parameter names (C2i and D2i, i = 1. Number) in order. Further, the analysis unit 11 acquires the value (X2) inserted into the column (X) by INSERT. Then, the analysis unit 11 performs the following calculation to obtain the value of α (S44).
α = (X1-X2) / (ΣC1i · D1i-ΣC2j · D2j)

  Next, the analysis unit 11 determines whether or not α is the same in all the operation logs (screen unit logs) (S45). If α is the same value (S45: YES), the analysis unit 11 analyzes that the values of the two screen parameters have been subjected to a predetermined operation, and have been INSERTed and propagated to the column (X). Therefore, the analysis unit 11 estimates that there has been indirect data propagation from the two screen parameters to the column (X) (S46), and loops the processing to the next combination. If α is not the same value (S45: NO), the analysis unit 11 does not presume that indirect data has propagated from the two screen parameters to the column (X), and proceeds to the next combination. Loop.

  FIG. 6 is a flowchart showing the process (S10) of "estimating the relationship between the two columns of the SELECT result and the INSERT value". As shown in FIG. 6, when the process of S10 is started, the analysis unit 11 searches the screen unit log for a SELECT to be executed before executing the INSERT for the column (X), and the search result is obtained. From the columns, a combination (C, D) in which two columns are selected is created (S50).

  Next, the analysis unit 11 executes a loop process from S51 to S56 for each created combination. When the loop processing is started, the analysis unit 11 selects one screen unit log, and retrieves a SELECT statement search result value corresponding to two column names (C1i and D1i, i = 1. The number of search results) is extracted, and the sum of products (ΣC1i · D1i) is calculated (S52). Further, the analysis unit 11 acquires the value (X1) inserted into the column (X) by INSERT.

Next, the analysis unit 11 retrieves a SELECT statement search result value corresponding to two column names (C2i and D2i, i = 1... N, respectively, where n is a parameter) for each screen unit log other than the one used in S52. ) Is taken out in order. Further, the value X2 inserted into the column (X) is acquired by INSERT. Then, the analysis unit 11 calculates the value of α by performing the following calculation (S53).
α = (X1-X2) / (ΣC1i · D1i-ΣC2j · D2j)

  Next, the analysis unit 11 determines whether or not α is the same in all operation logs (screen unit logs) (S54). If α is the same value (S54: YES), the analysis unit 11 analyzes that the values of the two columns (C, D) have been subjected to a predetermined operation and have been inserted into the column (X) and propagated. . Therefore, the analysis unit 11 estimates that there has been indirect data propagation from the two columns (C, D) to the column (X) (S55), and loops the processing to the next combination. If α is not the same value (S54: NO), the analysis unit 11 does not presume that indirect data transmission from the two columns (C, D) to the column (X), and Loop through the process to the combination.

  FIG. 7 is a flowchart illustrating the process (S11) of “estimating the relationship between two types of SELECT result columns and INSERT values associated via key values”. As shown in FIG. 7, when the processing in S11 is started, the analysis unit 11 selects one screen-based log and becomes a search target by a SELECT statement executed before the INSERT to be inserted into the column (X). A table with one or more results is searched (S60). The analysis unit 11 obtains a plurality of tables (table group) corresponding to the condition by searching the tables in S60.

  Next, the analysis unit 11 creates a combination of two different tables from the table group obtained in S60, and has a relationship in which the column values of one table are included in the column values of the other table. The table / column pair (K1, K2) is extracted (S61).

  Next, the analysis unit 11 executes a loop process of S62 to S69 for each extracted table / column pair (K1, K2). When the loop processing is started, the analysis unit 11 extracts, from the search results from the two tables of the table / column pair (K1, K2), the search results in which the respective values (column values) of the column pairs are the same. Obtained (S63).

  Next, the analysis unit 11 selects columns (C, D) other than the column pair (K1, K2) obtained in S63 one by one from each of the two tables (S64).

  Next, the analysis unit 11 obtains C1i, D1i, i = 1... From the values (C, D) of the two columns selected in S64 from the search results associated with K1 and K2. . m, where m is the number of search results, and the product sum ΣC1i · D1i is calculated. Further, the analysis unit 11 acquires the value (X1) inserted into the column (X) by INSERT (S65).

Next, the analysis unit 11 performs the same processing as that of S61 to S65 on the screen unit log other than the screen unit log selected in the first (S60), and inserts the product sum value ΣC2i · D2i and INSERT into the column (X). Obtained value (X2). Then, the analysis unit 11 performs the following calculation to obtain the value of α (S66).
α = (X1-X2) / (ΣC1i · D1i-ΣC2j · D2j)

  Next, the analysis unit 11 determines whether or not α is the same in all the operation logs (screen unit logs) (S67). When α is the same value (S67: YES), the analysis unit 11 analyzes that the values of the two columns (C, D) have been subjected to a predetermined operation and have been inserted into the column (X) and propagated. . Therefore, the analysis unit 11 estimates that there has been indirect data propagation from the two columns (C, D) to the column (X) (S68), and loops the process to the next column pair. If α is not the same value (S67: NO), the analyzing unit 11 does not estimate that there has been indirect data propagation from the two columns (C, D) to the column (X), and Loop through the process to column pairs.

  FIG. 8 is a flowchart showing the process (S12) of "estimating the relationship between the screen parameter associated with the key value, the column of the SELECT result, and the INSERT value". As shown in FIG. 8, when the process of S12 is started, the analysis unit 11 selects one screen-based log and becomes a search target by a SELECT sentence executed before the INSERT to be inserted into the column (X). A table (table group) having one or more results is obtained (S70).

  Next, the analysis unit 11 acquires, from the screen parameters, a set of parameter names that appear the same number of times, regards the set of parameter names that appear the same number of times as a column name, and considers the set of parameter values for each repetition as a search result of SELECT ( S71).

  Next, the analysis unit 11 creates a combination of the table group acquired in S70 and the set of the screen parameters acquired in S71 as a column name and a SELECT search result in a pseudo manner. Then, the analysis unit 11 extracts a table / screen parameter pair (K1, K2) in which the column value of the table is included in the screen parameter value (S72).

  The analysis unit 11 performs the process (S11) of “estimating the relationship between the column of two types of SELECT results and the INSERT value related via the key value” for the table / screen parameter pair (K1, K2) extracted in S72. The same processing as S64 to S68 is performed.

  FIG. 9 is a flowchart illustrating an operation example of the analysis support apparatus 1 according to the embodiment, and is an example of a process in the case of the above-described UPDATE. As shown in FIG. 9, the processing (S1 to S3) until a set of operation logs is created for each screen is performed in the same manner as the processing for INSERT (see FIG. 2).

  Next, the analysis unit 11 executes a loop process of S80 to S84 for each set of the created operation logs for each screen. When the loop processing is started, the analysis unit 11 extracts a combination of executing the UPDATE and executing the SELECT on the UPDATE target table before executing the UPDATE from the set of operation logs to be processed. The analysis unit 11 extracts a column (X) in which the column value of the SELECT result is different from the column value updated by UPDATE (S81).

  Next, the analysis unit 11 performs the process (S7 to S12) of “when a value that does not directly correspond is inserted by INSERT” in the column (X) extracted in S81 in the UPDATE target table, The process is executed by replacing the INSERT value with a value to be updated by UPDATE (S82).

  Next, the analysis unit 11 performs the process (S7 to S12) of “when a value that does not directly correspond is inserted by INSERT” in the column (X) extracted in S81 in the UPDATE target table, Instead of the INSERT value, it is replaced with (value to be updated by UPDATE)-(SELECT search result value) and executed (S83).

  When the above loop processing is completed, the output unit 13 outputs the analysis result of the analysis unit 11 by the loop processing on a display or by printing by a printer (S15).

  Here, the analysis processing in the analysis support apparatus 1 will be described with reference to a specific embodiment. In the embodiment described below, it is assumed that all columns in the table of the business DB to be analyzed are of numeric type.

  FIG. 10 is a diagram illustrating a table configuration of the business DB 21. As shown in FIG. 10, the business DB 21 is directed to an order receiving business, and includes a product master table T1, a detail table T2, an order receiving table T3, and a billing table T4. Note that, in the business DB 21, columns that are not relevant in the description are omitted with "...".

(Estimation of relationship between one column and INSERT value)
FIG. 11 is an explanatory diagram illustrating an example of the operation log (D11, D12, D13). As shown in FIG. 11, it is assumed that the operation log acquired from the business system 2 to be analyzed is the contents (values) of D11, D12, and D13. It is assumed that the operation logs D11, D12, and D13 are output in chronological order, and the execution order (seq) is assigned to each execution of the screen parameters and SQL (the execution time may be used).

  FIG. 12 is an explanatory diagram illustrating an example of the data tables (T11, T12, and T13). As shown in FIG. 12, the data table T11 is a table in which the IDs (urlNo), the executed URLs, and the seq for each screen parameter are collected from the operation logs D11, D12, and D13. Similarly, the data table T12 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID of the screen parameter corresponding to the SQL, and the seq. The data table T13 is a set of columns (SetC).

  According to the data table T11, in this embodiment, the screens are all the same, and the executed SQL (one SELECT, one INSERT) is the same except for the bind value. Is a set of

  When a column of the table to be inserted and a SELECT preceding the column are associated with each other by a known method, the value of the slip number column of the search result of the preceding SELECT always matches the value of the slip number of the subsequent INSERT. Can be detected. Therefore, in S5, the analysis support device 1 estimates that there is direct data propagation from the slip number column of the order receiving table T3 to the slip number column of the billing table T4.

  Regarding the billing amount column of the billing table T4 not related above, the column set SetC included in the SELECT result executed before the INSERT is (slip number, tax-excluded amount, customer ID). The analysis unit 11 of the analysis support device 1 performs a loop process of S4 to S14 for each of these sets SetC to analyze indirect data propagation.

  First, it is checked whether the slip number column and the billing amount column have a propagation relationship. From the data of seqNo = 1 in the data table T12, the slip number = 101 and the charged amount = 25340. Therefore, c1Val = 101 and x1Val = 25340. Next, similarly, c2Val = 102 and x2Val = 4928 from the data of seqNo = 2. When the value of α is obtained from these, α = (25340−4928) / (101−102) = − 20412.

  Similarly, from the data of seqNo = 3, c2Val = 103 and x2Val = 1580. When the value of α is calculated from these, α = (25340-1580) / (101-103) =-11880. Therefore, since α is not constant, it can be seen that this combination does not have a propagation relationship.

  Next, it is checked whether there is a propagation relationship between the tax-excluded amount column and the invoice amount column. From the “operation log 1”, the tax-excluded amount = c1Val = 23000 and the charged amount = x1Val = 25340. Next, c2Val = 4100 and x2Val = 4928 from “operation log 2”. When the value of α is calculated from these, α = (25340−4928) / (23000-4100) = 1.08.

  Similarly, from “operation log 3”, c2Val = 1000 and x2Val = 1580. When the value of α is obtained from these, α = (25340-1580) / (23000-1000) = 1.08. Therefore, α is constant, and β = x1Val−α · c1Val = 500. From the above, this combination has an indirect propagation relationship, and it can be estimated that the tax-excluded amount * 1.08 + 500 = the billing amount.

  In the combination of the customer ID column and the billing amount column, the customer ID = c1Val = 2001 and the billing amount = x1Val = 25340 from “Operation log 1”. Next, c2Val = 1507 and x2Val = 4928 from "operation log 2". When the value of α is obtained from these, α = (25340−4928) / (2001-1507) = 41.3. Similarly, c2Val = 3900 and x2Val = 1580 from “operation log 3”. When the value of α is obtained from these, α = (25340-1580) / (2001-3900) = − 12.5. Therefore, since α is not constant, it can be seen that this combination does not have a propagation relationship.

(Estimation of relationship between one screen parameter and INSERT value)
FIG. 13 is an explanatory diagram illustrating an example of the operation log (D21, D22, D23). As shown in FIG. 13, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D21, D22, and D23.

  FIG. 14 is an explanatory diagram illustrating an example of a data table (T21, T22, T23). As shown in FIG. 14, the data table T21 summarizes the ID (urlNo), the executed URL, and the seq for each screen parameter from the operation logs D21, D22, and D23. Similarly, the data table T22 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL. The data table T23 is a set of columns (SetC). Specifically, the set SetC obtained by extracting the HTTP parameters from the URLs of the operation logs D21, D22, and D23 is (ORDER, PRICE).

  First, it is checked whether or not there is a propagation relationship between the screen parameter ORDER and the billing amount column. Specifically, from urlNo = 1 in the data table T21, ORDER = c1Val = 101, and the charge amount = x1Val = 25340. Next, urlNo = 2 to c2Val = 102 and x2Val = 4928 in the data table T21. When the value of α is obtained from these, α = (25340−4928) / (101−102) = − 20412. Similarly, urlNo = 3 to c2Val = 103 and x2Val = 1580. When the value of α is calculated from these, α = (25340-1580) / (101-103) =-11880. Therefore, since α is not constant, it can be seen that this combination does not have a propagation relationship.

  Next, it is checked whether there is a propagation relationship between the screen parameter PRICE and the billing amount column. From urlNo = 1, PRICE = c1Val = 23000, and the charged amount = x1Val = 25340. Next, urlNo = 2 to c2Val = 4100 and x2Val = 4928. When the value of α is calculated from these, α = (25340−4928) / (23000-4100) = 1.08.

  Similarly, urlNo = 3 to c2Val = 1000 and x2Val = 1580. When the value of α is obtained from these, α = (25340-1580) / (23000-1000) = 1.08. Therefore, α is constant, and β = x1Val−α · c1Val = 500. From the above, this combination has an indirect propagation relationship, and it can be estimated that PRICE * 1.08 + 500 = charged amount.

(About the estimation of the relationship between the multiple repetition parameters of the screen parameter and the INSERT value)
FIG. 15 is an explanatory diagram illustrating an example of the operation log (D31, D32, D33). As shown in FIG. 15, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D31, D32, and D33.

  FIG. 16 is an explanatory diagram illustrating an example of the data tables (T31, T32, and T33). As shown in FIG. 16, the data table T31 summarizes the ID (urlNo), the executed URL, and the seq for each screen parameter from the operation logs D31, D32, and D33. Similarly, the data table T32 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL.

  It is assumed that the variable name of the screen parameter (for example, CODE-1) is composed of a character string in which a numerical value (for example, 1) is connected after a non-numeric value (for example, CODE-). To separate. Then, by forming a set for each of the same numeric character strings, a set of parameters shown in the data table T33 is created from the URL value of the operation log D31. Specifically, a set of two parameter names is (CODE-, UNITPRICE-), (CODE-, QTY-), and (UNITPRICE-, QTY-). The analysis unit 11 estimates whether or not there is an indirect propagation relationship with the tax-excluded amount inserted in INSERT by the following processing.

  In the case of the set (CODE-, UNITPRICE-), the values of two parameters are C = (123, 130) and D = (1000, 1500) from urlNo = 1 in the data table T33. Therefore, the product sum value is ΔC1i · D1i = 123 * 1000 + 130 * 1500 = 318000, and X1 = 23000.

  When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 3280000, and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (318000−3280000) = − 6.4 * 10−3.

  Similarly calculated with urlNo = 3, α = (23000-1580) / (318000-60100) = 8.3 * 10−2. Therefore, since α is different from each other, it is determined that no data is propagated in this combination.

  Similarly, in the case of the set (CODE-, QTY-), the product sum value ΣC1i · D1i = 123 * 20 + 130 * 2 = 2720 and X1 = 23000 from urlNo = 1 in the data table T33. When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 800 and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000−4100) / (2720−800) = 9.8.

  Similarly calculated with urlNo = 3, α = (23000-1580) / (2720-8005) = − 4.1. Therefore, since α is different from each other, it is determined that no data is propagated in this combination.

  Similarly, in the case of the set (UNITPRICE-, QTY-), the product sum value ΣC1i · D1i = 1000 * 20 + 1500 * 2 = 23000 and X1 = 23000 from urlNo = 1 in the data table T33. When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 4100 and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (23000-4100) = 1.

  Similarly, when urlNo = 3, α = (23000-1580) / (23000-1580) = 1. Therefore, since α coincides with each other, it is presumed that UNITPRICE- * and QTY- * have an indirect propagation relationship in the tax-excluded amount column of the order receiving table T3.

(About estimating the relationship between the two columns of the SELECT result and the INSERT value)
FIG. 17 is an explanatory diagram illustrating an example of the operation log (D41, D42, D43). As shown in FIG. 17, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D41, D42, and D43.

  FIG. 18 is an explanatory diagram illustrating an example of a data table (T41, T42, T43). As shown in FIG. 18, the data table T41 summarizes the ID (urlNo), the executed URL, and the seq for each screen parameter from the operation logs D41, D42, and D43. Similarly, the data table T43 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL. The data table T42 is a collection of columns (detail ID, slip number, product ID, unit price excluding tax, quantity) of the detail table T2 searched by SELECT prior to INSERT.

  The analysis unit 11 takes out two columns (specification ID, slip number, product ID, unit price excluding tax, quantity) from the details table T2 searched by SELECT prior to INSERT, creates a set, and sets a tax-excluded amount to be inserted by INSERT. Whether or not there is an indirect propagation relation is estimated by the following processing.

  For example, in the case of a set (product ID, unit price without tax), the values of two columns from urlNo = 1 in the data table T42 are C = (123, 130) and D = (1000, 1500). Therefore, the sum of products ΣC1i · D1i = 123 * 1000 + 130 * 1500 = 318000, and X1 = 23000.

  When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 3280000, and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (318000−3280000) = − 6.4 * 10−3.

  Similarly calculated with urlNo = 3, α = (23000-1580) / (318000-60100) = 8.3 * 10−2. Therefore, since α is different from each other, it is determined that no data is propagated in this combination.

  Similarly, in the case of (tax-excluded unit price, quantity), the product sum value ΣC1i · D1i = 1000 * 20 + 1500 * 2 = 23000 and X1 = 23000 from urlNo = 1 in the data table T42. When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 4100 and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (23000-4100) = 1.

  Similarly, when urlNo = 3, α = (23000-1580) / (23000-1580) = 1. Therefore, since α coincides, it is estimated that there is an indirect propagation relationship between the tax-excluded unit price column and the quantity column of the detail table T2 and the tax-excluded amount column of the order receiving table T3.

(Estimation of relationship between INSERT value and columns of two types of SELECT results related via key values)
FIG. 19 is an explanatory diagram illustrating an example of the operation log (D51, D52, D53). As shown in FIG. 19, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D51, D52, and D53.

  FIG. 20 is an explanatory diagram illustrating an example of the data tables (T51, T52, T53, and T54). As shown in FIG. 20, the data table T51 summarizes the ID (urlNo), the executed URL, and the seq for each screen parameter from the operation logs D51, D52, and D53. Similarly, the data table T52 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL.

  When the tax-excluded amount column of the order receiving table T3 inserted by INSERT is set to column X, in the operation logs D51, D52, and D53 of this embodiment, two types of SELECT prior to INSERT are executed in each screen unit log. For these two types, the data table T53 summarizes the search results from the detail table T2. The data table T54 summarizes the search results from the product master table T1.

  The analysis unit 11 selects one column from each of the data table T53 and the data table T54, and searches for a column in which one column value includes the other column value. Thereby, the analysis unit 11 can detect that the product ID of the data table T53 and the product ID of the data table T54 correspond to the column pair (K1, K2). Therefore, the analysis unit 11 selects columns other than the column pair (K1, K2) one by one from the data table T53 and the data table T54.

  When the unit price excluding the tax in the product master table T1 is selected as the column C and the quantity in the detail table T2 is selected as the column D, when the product sum value is obtained using urlNo = 1 and the product ID having the same value, {C1i · D1i = 1000 * 20 + 1500 * 2 = 23000. In this case, X1 = 23000.

  When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 4100 and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (23000-4100) = 1.

  Similarly, when urlNo = 3, α = (23000-1580) / (23000-1580) = 1. Therefore, since α matches, the tax-excluded unit price column of the product master table T1 and the quantity column of the detail table T2 are indirectly propagated to the tax-excluded amount column of the order receiving table T3 using the respective product ID columns as keys. Presumed to be relevant.

(Regarding the estimation of the relationship between the screen parameter associated with the key value, the column of the SELECT result, and the INSERT value)
FIG. 21 is an explanatory diagram illustrating an example of the operation log (D61, D62, D63). As shown in FIG. 21, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D61, D62, and D63.

  FIG. 22 is an explanatory diagram illustrating an example of the data tables (T61, T62, T63, and T64). As shown in FIG. 22, the data table T61 summarizes IDs (urlNo), executed URLs, and seq for each screen parameter from the operation logs D61, D62, and D63. Similarly, the data table T62 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL.

  As the screen parameters of the present embodiment, there are two types of repetition of parameters (for example, CODE- * and QTY- *) as shown in "Regarding Estimation of Relationship Between Plural Repetition Parameters of Screen Parameters and INSERT Value". The data table T63 summarizes the extracted repeated values. When the tax-excluded amount column of the order receiving table T3 inserted by INSERT is set to column X, in the operation log of the present embodiment, one type of SELECT prior to INSERT is executed in each screen unit log. The data table T64 summarizes search results from the product master table T1.

  The analysis unit 11 selects one column from each of the data table T63 and the data table T64, and searches for a column in which one column value includes the other column value. Accordingly, the analysis unit 11 can detect that the CODE- of the data table T63 and the product ID of the data table T64 correspond to the column pair (K1, K2). Therefore, the analysis unit 11 selects columns other than the column pair (K1, K2) one by one from the data table T63 and the data table T64.

  When the unit price excluding tax of the product master table T1 is selected as the column C, and the screen parameter QTY- is selected as the column D, when urlNo = 1, the product ID and CODE- are used to obtain a product-sum value using the same value pair. ΣC1i · D1i = 1000 * 20 + 1500 * 2 = 23000. In this case, X1 = 23000.

  When similarly calculated with urlNo = 2, the product sum value ΣC2j · D2j = 4100 and X2 = 4100. Therefore, when the value of α is obtained, α = (X1−X2) / (ΣC1i · D1i−ΣC2j · D2j) = (23000-4100) / (23000-4100) = 1.

  Similarly, when urlNo = 3, α = (23000-1580) / (23000-1580) = 1. Therefore, since α coincides with each other, the unit price column excluding tax in the product master table T1 and the QTY- * of the screen parameter are compared with the column excluding tax in the order receiving table T3 using the respective product ID columns and CODE- * as keys. It is estimated that there is an indirect propagation relationship.

(Relation estimation when updating values that cannot be directly associated with UPDATE)
FIG. 23 is an explanatory diagram illustrating an example of the operation log (D71, D72, D73). As shown in FIG. 23, it is assumed that the operation logs acquired from the business system 2 to be analyzed are the contents (values) of D71, D72, and D73.

  FIG. 24 is an explanatory diagram illustrating an example of the data tables (T71, T72, and T73). As shown in FIG. 24, the data table T71 summarizes IDs (urlNo), executed URLs, and seq for each screen parameter from the operation logs D71, D72, and D73. Similarly, the data table T72 summarizes the ID (sqlNo) for each SQL, the executed SQL, the bind value and the search result at that time, the ID and seq of the screen parameter corresponding to the SQL. The data table T73 summarizes the updated value of the bill and the difference between the bills.

  When the billing amount column of the billing table T4 is the column X, the column set SetC included in the SELECT result executed before the INSERT is the slip number of the order receiving table, the tax-excluded amount, the customer ID, and the slip number of the billing table.

  As for “association with one column”, a case where it is checked whether or not there is a propagation relationship between the tax exemption amount column and the invoice amount column is illustrated. First, the value itself to be updated by UPDATE is processed as the value of column X.

  From the “operation log 1”, the tax-excluded amount = c1Val = 23000 and the charged amount = x1Val = 25340. Next, c2Val = 4100 and x2Val = 5428 from “operation log 2”. From these, α = (25340-5428) / (23000-4100) = 1.05. Similarly, c2Val = 1000 and x2Val = 2580 from “operation log 3”. When the value of α is determined from these, α = (25340−2580) / (23000−1000) = 1.03, and α is not constant.

  Next, processing is performed using a difference value (a difference between billing amounts in the data table T73) with a search result of SELECT executed prior to UPDATE. From the “operation log 1”, the tax-excluded amount = c1Val = 23000, and the difference between the charged amounts = x1Val = 24840.

  Next, c2Val = 4100 and x2Val = 4428 from “operation log 2”. When the value of α is calculated from these, α = (24840-4428) / (23000-4100) = 1.08.

  Similarly, c2Val = 1000 and x2Val = 1080 from “operation log 3”. When the value of α is calculated from these, α = (24840-1080) / (23000-1000) = 1.08. Therefore, α is constant, and β = x1Val−α · c1Val = 0. From the above, it is estimated that this combination has an indirect propagation relationship, and that the tax-excluded amount * 1.08 = the updated value of the billing amount.

  In the above embodiment, the case of performing the “association with one column” process is exemplified. However, by performing the same in other cases, the UPDATE update value or the difference between the update values and the indirect data propagation It can be estimated whether or not they are related.

  FIG. 25 is an explanatory diagram illustrating an example of the display screen G. As shown in FIG. 25, the output unit 13 displays the analysis result of the analysis unit 11 on the display screen G of the display. Specifically, the output unit 13 displays, on the display screen G, table / column images G1 to G5 listing the columns of the business DB 21 and the variable values in the screen parameters. Then, based on the analysis result of the analysis unit 11, the output unit 13 performs connection relationship displays G11 and G12 in which columns and variable values having a corresponding relationship are connected by a line.

  For example, a column or variable value analyzed as having direct data propagation is a connection relationship display G11 such as a solid line. The columns and variable values analyzed as having indirect data propagation are indicated by a connection relationship display G12 such as a broken line. Thus, the user can easily distinguish between the analysis result analyzed as having direct data propagation and the analysis result analyzed as having indirect data propagation.

  As described above, the analysis support apparatus 1 determines at least one of the input parameters such as the screen parameters and the bind variable values, and the reading of data from the business DB and the writing of data to the business DB based on the input parameters. A plurality of operation logs including an access record in which data has been written are acquired. Then, the analysis support apparatus 1 performs the operation in the value included in the input parameter or the access record in each of the operation logs including the access record related to the predetermined column of the business DB in the acquired plurality of operation logs. A predetermined operation is performed by combining the value read from the business DB with the value written to the business DB in the access record. When the calculation result (α) is constant, the analysis support apparatus 1 analyzes that there is data propagation in the column, and outputs the analyzed column when there is propagation. For this reason, for example, the analysis support apparatus 1 calculates the value read from the business DB in the method, and also grasps indirectly related data propagation such as writing the value after the calculation to another column. be able to. The analysis support device 1 is configured to acquire the operation log collected as the DB access data 20 by the business system 2. Thereby, for example, the data propagation around the business DB of the business system 2 can be grasped in the upstream process of maintenance development, and there is an advantage that the influence of the system change on the current system can be more easily investigated.

  In addition, each component of each illustrated device does not necessarily need to be physically configured as illustrated. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or a part thereof may be functionally or physically distributed / arbitrarily divided into arbitrary units according to various loads and usage conditions. Can be integrated and configured. For example, the analysis unit 11, the storage unit 12, or the output unit 13 may be connected via a network as an external device of the analysis support device 1. Alternatively, the analysis unit 11, the storage unit 12, or the output unit 13 may be included in another device and connected to a network to cooperate to realize the function of the analysis support device 1.

  The various processes described in the above embodiments can be realized by executing a prepared program on a computer such as a personal computer or a workstation. Therefore, an example of a computer that executes an analysis support program having the same functions as the above embodiment will be described below with reference to FIG.

  FIG. 26 is an explanatory diagram illustrating an example of the computer 200 that executes the analysis support program 270a. As shown in FIG. 26, the computer 200 includes an operation unit 210a, a speaker 210b, a camera 210c, a display 220, and a communication unit 230. Further, the computer 200 has a CPU 250, a ROM 260, an HDD 270, and a RAM 280. These units 210 to 280 are connected via a bus 240.

  The HDD 270 stores in advance an analysis support program 270a that performs the same function as the analysis unit 11, the storage unit 12, and the output unit 13 described in the above embodiment. The analysis support program 270a may be appropriately integrated or separated, similarly to the respective components of the analysis unit 11, the storage unit 12, and the output unit 13. For example, all data stored in the HDD 270 need not always be stored in the HDD 270, and only data necessary for processing need be stored in the HDD 270.

  In the computer 200, the CPU 250 reads the analysis support program 270a from the HDD 270 and expands the analysis support program 270a in the RAM 280. Thus, the analysis support program 270a functions as the analysis support process 280a. The analysis support process 280a expands various data read from the HDD 270 into an area allocated to itself on the RAM 280 as appropriate, and executes various processes based on the expanded various data. Note that the analysis support process 280a includes processing executed by the analysis unit 11, the storage unit 12, and the output unit 13. In addition, all the processing units virtually implemented on the CPU 250 need not always operate on the CPU 250, and only the processing units required for processing need be virtually implemented.

  The analysis support program 270a need not always be stored in the HDD 270 or the ROM 260 from the beginning. For example, each program is stored in a “portable physical medium” such as a flexible disk inserted in the computer 200, so-called FD, CD-ROM, DVD disk, magneto-optical disk, IC card, or the like. Then, the computer 200 may acquire and execute each program from these portable physical media. Further, each program may be stored in another computer or a server device connected to the computer 200 via a public line, the Internet, a LAN, a WAN, or the like, and the computer 200 may acquire and execute each program from these. It may be.

(Another embodiment)
Another embodiment of the above-described embodiment will be described. In the business system 2, in the case of an order receiving business as an example, a plurality of item items included in one order are input on a screen, and processing may be performed for each group. For example, a product code and a set of ordered quantities are input as item items from the screen, and the input data is used to write the product code and ordered value for each item in the item table. At the same time, there may be a case where a value obtained by calculating the total amount of money in the detail unit is written in the detail table. Further, the order table is updated using the order slip itself.

  The SQL executed upon input from the screen is a mixture of two types of processing: the processing of the order receipt itself and the processing repeated for each item. Therefore, in the above-described embodiment, it may be difficult to estimate the correspondence between data.

  FIG. 28 is an explanatory diagram illustrating an example of data propagation in an order receiving business. FIG. 29 is an explanatory diagram illustrating a case where detection of data propagation has failed. In FIG. 28, an operation log D81 is a log of a screen input (HTTP) in an order receiving operation. The operation log D82 is an SQL log that is executed in response to a screen input. Further, a product master table T1, a detail table T2, an order receiving table T3, and a message screen parameter table T5 indicate data tables related to the order receiving business, and solid lines and broken lines between the tables indicate data propagation between items in the table. Note that a solid line indicates direct data propagation, and a broken line indicates indirect data propagation. In the example of FIG. 29, the detection of the data propagation relationship in the detail table T2 has failed.

  As shown in FIG. 28, in the operation logs D81 and D82 in the order receiving business, there is a data propagation relation to the order table T3 and a data propagation relation to the detail table T2. In the example of the operation logs D81 and D82, the SQL regarding the two items is executed. It corresponds to two different values of "10" and "21". Therefore, in the above-described embodiment, it is difficult to estimate the correspondence, and as shown in FIG. 29, the detection of the data propagation relationship in the detail table T2 fails.

  For this reason, in another embodiment, the analysis unit 11 acquires a key candidate of a column or a parameter of a screen, which is a key, for each screen-based operation log (screen-based log) generated by input from one screen. Next, the analysis unit 11 reads out from the business DB 21 in an access record in which a value or data included in an input parameter such as a screen parameter or a bind variable value is written for each column having a key candidate (key column candidate). The calculated value is combined with the value written to the business DB 21 in the access record, and a predetermined calculation is performed in the same manner as in the above-described embodiment. Thus, in another embodiment, even in the case where it is difficult to estimate the correspondence of data in the above-described embodiment (for example, data propagation for each specification), data propagation can be estimated.

  Hereinafter, another embodiment will be described in detail. In another embodiment described below, an application executed in the business system 2 to be analyzed is executed online, an internal process of the business application is executed using a value input from a certain screen, and the screen is stateless on a screen basis. It is described on the assumption that: However, this premise is an example, and the present invention can be applied to batch processing and the like. Also, it is assumed that the SQL statement is described using a bind variable, but a part corresponding to the bind variable is extracted from the SQL in which the variable value is embedded and converted into the SQL, the bind variable, and the bind variable value. It is possible to do.

  The operation log acquisition unit 10 accesses a predetermined log file in the business system 2 to be analyzed, and obtains a set of operation logs (screen unit log) in the business process triggered by input from one screen of the terminal device. collect. More specifically, the operation log acquisition unit 10 transmits HTTP parameters (screen parameters) input on the screen and input to the business application, and a series of SQL logs (SQL statements, input parameters (binding parameters) executed when the screen is triggered. (Variable value), in the case of a SELECT statement, a search result) and information (time, etc.) that can determine their execution order.

  The analysis unit 11 searches the screen unit log for each screen log acquired by the operation log acquisition unit 10 for a value (key value) for distinguishing details and a screen parameter or column having the key value. Here, since the key value and the screen parameter name and the column name having the key value are unknown, the analysis unit 11 first sets the value appearing in the data transition destination table (the table handled by the INSERT or UPDATE statement in SQL) as the key. Assuming a value, generate a column candidate as a key.

  Next, the analysis unit 11 groups the SQLs having the same key value as a group, performs the same operation as the above-described embodiment using the SQLs belonging to the group and the screen parameters, and performs data propagation (direct data propagation and indirect data propagation). Data propagation) is detected.

Specifically, the analysis unit 11 executes the following calculations (1) to (3).
(1) The same value as the key value candidate at the data transition destination is searched from the data transition source to generate a column or screen parameter candidate as a key.
(2) SQL candidates having the same key candidate are grouped.
(3) The same calculation as in the above-described embodiment is performed on the above SQL group to detect data propagation.

  FIG. 30 is a flowchart illustrating an operation example of the analysis support device 1 according to another embodiment. The same processes as those in the above-described embodiment are denoted by the same reference numerals, and description thereof will be omitted.

  As illustrated in FIG. 30, the analysis unit 11 executes a first loop process of S4a to S14a for each set of the created screen unit logs. Specifically, when the first loop process is started, the analysis unit 11 acquires a key value candidate for each screen unit log (S100). Key value candidates include values set in INSERT or UPDATE included in the screen unit log, values appearing in the WHERE clause of UPDATE, and the like.

  Next, the analysis unit 11 creates an SQL sequence (sentence) having a key value for each candidate key value (S101). The SQL column referred to here is a SELECT (0 or more) having one INSERT or UPDATE including a key value and a key executed before the one INSERT or UPDATE in the WHERE clause in the SQL column of the screen unit log. ).

  Next, the analysis unit 11 creates a column name (key column candidate) having a key value by INSERT or UPDATE in the SQL string created in S101 (S102). Specifically, the analysis unit 11 searches the SQL string created in S101 for a table / column (data transmission destination) to be INSERTed or UPDATEd, and creates a key column candidate.

  Next, the analysis unit 11 creates a list having key value candidates, SQL columns, and key column candidates based on the results of S100, S101, and S102 (S103). Next, the analysis unit 11 executes the second loop processing of S104 to S109 for each key column candidate based on the list created in S103.

  When the second loop process is started, the analysis unit 11 performs the same operation as in the above-described embodiment on the SQL (column) corresponding to the key column candidate, and detects data propagation (S105). Thus, the analysis unit 11 detects data propagation from the table.

  Next, the analysis unit 11 calculates the key value (K) of the screen parameter (S106). Specifically, the analysis unit 11 uses a group in which a part of the screen parameter name is common (for example, in the case of “code-1”, “code-2”, and “code-3”, “code-” is common). Is searched for a group having one key value as a group value and different values in all other groups. Then, the analysis unit 11 obtains the number of the key value in the found group (K-th).

  Next, the analysis unit 11 extracts the K-th screen parameter (S107). Specifically, another group having the same number of key values as the group in S106 is searched, and the K-th parameter is extracted (a set of screen parameters corresponding to the key value). However, a common character string (for example, “code-” for “code-1”, “code-2”, and “code-3”) is used for the parameter name.

  Next, the analysis unit 11 combines the K-th screen parameter extracted in S107 corresponding to the key value and the SQL sequence corresponding to the same key value, performs the same operation as in the above-described embodiment, and performs data propagation. Detection is performed (S108). Thereby, the analysis unit 11 detects data propagation from the screen parameters.

  Here, another embodiment of the analysis support apparatus 1 will be described with reference to a more specific specific example. In the specific example described below, it is assumed that the columns of the table of the business DB 21 to be analyzed are all numeric type. However, even if a character string column contains a numeric value, it can be processed in the same way as a numeric type by appropriately converting the character string and treating it as a numeric value.

  FIG. 31 is an explanatory diagram illustrating an example of an operation log. As shown in FIG. 31, the operation logs D91 and D92 include an SQL execution statement (SQL and bind value, and in the case of SELECT, a value obtained by a search) executed by the screen operation indicated by the URL. Have. The operation logs D91 and D92 execute the same type of SQL statement when the screen operation is triggered (for example, although the bind value is different, the SELECT statement to be executed is the same as the INSERT statement). Belong.

  In S100, the analysis unit 11 acquires key value candidates for each of the operation logs D91 and D92. FIG. 32 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 32 illustrates a data table T81 obtained by acquiring key value candidates for each of the operation logs D91 and D92.

  As illustrated in FIG. 32, the analysis unit 11 acquires, as a key value candidate, a value set by an INSERT statement for each screen log (operation logs D91 and D92). In this example, since the UPDATE was not executed, the value set in the UPDATE and the value where the WHERE clause appears are not obtained.

  Next, in S103, the analysis unit 11 creates a list of key value candidates of the screen unit log, an SQL column including the key value candidates, and a key column candidate having a key value in INSERT or UPDATE.

  FIG. 33 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 33 is a diagram illustrating a data table T82a indicating a list of key column candidates in the operation log D91 and a data table T82b indicating a list of key column candidates in the operation log D92.

  As shown in the data table T82a in FIG. 33, for example, the INSERT statement having the key value candidate “001” in the operation log D91 and the SELECT statement having the key value in the WHERE clause are (SQL1, SQL3), and the key is the INSERT statement in the INSERT statement. The column in which the value candidate “001” appears is “specification. Product code”.

  FIG. 34 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 34 is a diagram illustrating a data table T83 showing a SQL column list for each key column candidate using the data tables T82a and T82b of FIG. is there. As shown in FIG. 34, the analysis unit 11 performs the same calculation as in the above-described embodiment using the SQL sequence corresponding to the key column candidate for each key column candidate in the data table T83, and detects data propagation.

  In the illustrated example, since only INSERT exists in the key column candidates other than “specification. Merchandise code”, no data propagation occurs. On the other hand, when the key column candidate is “specification. Product code”, each SQL column is a combination of SELECT and INSERT, and data propagation is detected.

  FIG. 35 is an explanatory diagram illustrating detection of data propagation. Specifically, as shown in FIG. 35, data propagation from “product master. Product code” to “specification. Product code” is detected, and a data propagation relationship between the product master table T1 and the specification table T2 is detected. You.

  Next, the analysis unit 11 creates a group having a part of the screen parameter name in common. FIG. 36 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 36 illustrates a data table T84 obtained by creating a group having a part of screen parameter names in common from the operation logs D91 and D92. FIG. In this example, the "product code-" of "product code-1, product code-2" and the "order quantity-" of "order quantity-1, order quantity-2" correspond to each other, as shown in FIG. Then, a data table T84 is obtained.

  Next, for each key column candidate, the analysis unit 11 checks whether or not the key value is included in the screen parameter.

  FIG. 37 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 37 is a diagram illustrating a data table T85 obtained by checking whether a key value is included in a screen parameter for each key column candidate. It is. As shown in FIG. 37, for example, in the “operation log 1”, the key value “020” of the key column candidate “specification. Product code” exists in the second value (K = 2) of “product code −”. A value “product code− = 020 & order quantity− = 21” combined with “21” which is the second value of the group other than “product code−” (“order quantity−” in this example) is obtained as a screen parameter. . Note that the screen parameters can be similarly obtained for the other operation logs and key values of the key column candidate “detail.commodity code”.

  FIG. 38 is an explanatory diagram illustrating an example of a data table. More specifically, FIG. 38 is a diagram illustrating a data table T86 having a combination of an operation log, a key value, a screen parameter, and an SQL in which a corresponding screen parameter exists. It is. The analysis unit 11 obtains a data table T86 shown in FIG. 38 based on the data table T85. The analysis unit 11 performs the same operation as in the above-described embodiment using the data table T86, and detects data propagation.

FIG. 39 is an explanatory diagram illustrating detection of data propagation. Specifically, as shown in FIG. 39, when data propagation is detected using the data table T86, the following relationship is detected.
"Product master. Product code" → "Details. Product code"
"(Screen) Product code-" → "Details. Product code"
"(Screen) Order quantity-" → "Details. Order quantity"
"(Screen) Order quantity-* Product master. Unit price" → "Details. Total amount"

  As a result, the data propagation relationship of the product master table T1, the detail table T2, and the screen parameter table T5 is detected. Therefore, in another embodiment, it is possible to detect the data propagation relationship at the location where the detection of the data propagation relationship has failed in FIG.

DESCRIPTION OF SYMBOLS 1 ... Analysis support apparatus 2 ... Business system 10 ... Operation log acquisition part 11 ... Analysis part 12 ... Storage part 13 ... Output part 20 ... DB access data 21 ... Business DB
200 Computer T1 Merchandise master table T2 Detail table T3 Order table T4 Billing tables T11 to T86 Data tables D11 to D92 Operation log G Display screens G1 to G5 Table / column images G11, G12 Connection relationships display

Claims (6)

Computer
An input parameter and a plurality of operation logs including an access record in which at least data is written out of reading data from the database and writing data to the database based on the input parameter,
In the obtained plurality of operation logs, in each of the operation logs including the access record according to the predetermined column of the database, the value included in the input parameter or read from the database in the access record. If the result of performing a predetermined operation in combination with the value and the value written to the database in the access record is constant, analyze that there is data propagation in the column,
An analysis supporting method, characterized by executing a process of outputting a column analyzed when there is propagation. The analyzing process includes integrating, adding, or multiplying and accumulating a value included in the input parameter or a value read from the database in the access record and a value written to the database in the access record. 2. The analysis support method according to claim 1, wherein when the coefficient relating to the calculation is constant, it is analyzed that there is data propagation in the column. In the analyzing process, for an update related to a predetermined column of the database, a value included in the input parameter or a value read from the database in the access record, and writing to the database in the access record is performed. 3. The analysis support method according to claim 1, wherein when a difference from the calculated value is constant, it is analyzed that there is data propagation in the column. 4. The analyzing process obtains a key candidate of a column or a screen parameter as a key for each screen-based operation log generated by an input from one screen, and for each column having the key candidate, the input parameter The predetermined calculation is performed by combining a value included in the access record or a value read from the database in the access record, and a value written to the database in the access record. 4. The analysis support method according to any one of 3. On the computer,
An input parameter and a plurality of operation logs including an access record in which at least data is written out of reading data from the database and writing data to the database based on the input parameter,
In the obtained plurality of operation logs, in each of the operation logs including the access record according to the predetermined column of the database, the value included in the input parameter or read from the database in the access record. If the result of performing a predetermined operation in combination with the value and the value written to the database in the access record is constant, analyze that there is data propagation in the column,
An analysis support program for executing a process of outputting a column analyzed when the data is transmitted. An acquisition unit that acquires a plurality of operation logs including an input parameter and an access record in which at least data is written out of reading data from the database and writing data to the database based on the input parameter,
In the obtained plurality of operation logs, in each of the operation logs including the access record according to the predetermined column of the database, the value included in the input parameter or read from the database in the access record. An analysis unit that analyzes that the value and the result of performing a predetermined operation in combination with the value written to the database in the access record are constant, and that there is data propagation in the column.
An output unit that outputs a column analyzed when the data is transmitted.

Images