CN104750727B - A kind of column memory storage inquiry unit and column memory storage querying method - Google Patents
A kind of column memory storage inquiry unit and column memory storage querying method Download PDFInfo
- Publication number
- CN104750727B CN104750727B CN201310744231.9A CN201310744231A CN104750727B CN 104750727 B CN104750727 B CN 104750727B CN 201310744231 A CN201310744231 A CN 201310744231A CN 104750727 B CN104750727 B CN 104750727B
- Authority
- CN
- China
- Prior art keywords
- data
- cascade
- querying condition
- column
- storage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000005055 memory storage Effects 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000008520 organization Effects 0.000 claims abstract description 72
- 239000003550 marker Substances 0.000 claims abstract description 68
- 238000007689 inspection Methods 0.000 claims description 22
- 230000008859 change Effects 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 16
- 238000011835 investigation Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 8
- 238000013500 data storage Methods 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000001976 improved effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000003032 molecular docking Methods 0.000 description 1
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention discloses a kind of column memory storage inquiry unit, which is characterized in that described device includes: memory storage unit, for creating each column data in columns group storage line storage organization in memory；Data load units, for the target column data in the line storage organization to be loaded into the memory storage unit；Query engine, cascade querying condition for being inputted according to data consumer, the column data for being loaded into the memory storage unit to the data load units is inquired, and is obtained the column data station location marker for meeting the cascade querying condition, is sent to row reading unit；Row reading unit obtains line number and the consistent All Datarows information of the column data station location marker according to the column data station location marker that the query engine obtains in the line storage organization, is sent to data consumer.The simple storage in memory can be realized through the invention, and efficiently inquiry.The present invention also provides a kind of column memory storage querying methods.
Description
Technical field
Field of data storage of the present invention, more particularly to a kind of column memory storage inquiry unit and method.
Background technique
With the rapid development of Internet, various businesses data volume also sharply increases, it is direct for the storage mode of data
The efficiency of data query is influenced, and the efficiency of data query will have a direct impact on the treatment effeciency of various businesses, scientific progress
Data storage, guarantees the efficient read-write of data, is the basis for improving business processing efficiency.
It is stored in memory using which type of data structure, there is important meaning to EMS memory occupation and operation efficiency
Justice, due to memory resource limitation, by occupy resource it is lesser in a manner of stored, should answer first be it is how small just small,
As soon as int occupies 4 bytes in memory, N number of int should be minimum close to 4*N, and such memory consumption may be considered small
, if also smaller than this, can only be using compression and specific coding mode, and compress and coding will necessarily in operation into
The process that row is decompressed and decrypted by, this necessarily will affect efficiency, so memory table is in addition to special situation, it should use original text
Storage avoids unnecessary operation from consuming.
Current various storage methods, the most commonly used is by row storage, the storage unit of every row data knot having the same
Structure, will support more data type datas of storage multiple row in a line, it is most apply be inserted into data and access according to when be
With behavior unit, so the data of the overwhelming majority are to store by row, however inquire and often occur on column, when inquiry, is needed
The row pointer to object that every a line gets every a line is traversed, then gets the value of respective column by row pointer to object, if using general
Structure may be indexed value according to the index or title of column, and each column object unit lattice are compared and are looked into after going
It looks for, judges whether it is selected row, be put into result queue if it is line pointer.Value task of this execution, needs
Searching value and type conversion is carried out to every row or value compares in every row, can occupy a large amount of memory sources, reduce inquiry value
Efficiency.
The storage of prior art internal storage data is mostly using the storage of object line and complicated data structure storage, operation effect
Rate is not high, occupies a large amount of memory sources；It is to be stored using memory database, but equally consume in a large amount of there are also a kind of thinking
It deposits, for performance dependent on index, the interface used is complex, using SQL expression and resolver parsing inquiry.These technologies
Feature has been doomed its search efficiency and loading efficiency is not high, and condition flexibility is not strong, has more complicated API, is not suitable for making
Used in for the application scenarios efficiently inquired in memory.
Therefore a kind of storage inquiry mode for supporting high-speed and high-efficiency inquiry in memory is needed.
Summary of the invention
The present invention provides a kind of column memory storage inquiry units, which is characterized in that described device includes:
Memory storage unit, for creating each column data in columns group storage line storage organization in memory；
Data load units, for the target column data in the line storage organization to be loaded into the memory storage list
Member；
Query engine, the cascade querying condition for being inputted according to data consumer load the data load units
The column data for entering the memory storage unit is inquired, and the column data station location marker for meeting the cascade querying condition is obtained,
It is sent to row reading unit；
Row reading unit obtains in the line storage organization according to the column data station location marker that the query engine obtains
Line number and the consistent All Datarows information of the column data station location marker are taken, data consumer is sent to.
Preferably, described device further include:
Cache unit, for storing the inquiry data of the query engine, when the grade joint investigation of data consumer input
Inquiry condition is consistent with the cascade querying condition of storage, and when the corresponding columns group of the cascade querying condition does not change, by institute
The column data station location marker for meeting the cascade querying condition for stating storage is sent to row reading unit；
Period control unit deletes for storing the period for the memory storage unit set memory and exceeds institute in array
The column data for stating the memory storage period is also used to set caching period for the cache unit, deletes and exceed institute in cache unit
State the data cached of caching period；
Control unit is locked, for setting read lock to the memory storage unit and writing lock, controls the memory storage list
The read-write locking and unlocking of member.
Detailed, the memory storage unit further comprises:
Column data memory module is deposited with the columns group in line for creating the column data of column storage of array different lines
The field name of respective column names the columns group in storage structure, and the array index is column data station location marker；
Column data memory module storage is belonged to identical line for creating tables of data by tables of data memory module
The columns group of storage organization is stored in identical tables of data；
The memory storage unit may include multiple tables of data memory modules, and each tables of data memory module may include multiple
Column data memory module, each column data memory module save the data of a column in line storage organization, the column data
Station location marker is identical with the line number in its corresponding line storage organization；
The column data is preferentially stored in column data memory module using simple types.
Preferably, version number is arranged in the tables of data to save in each tables of data memory module, when the tables of data
When the corresponding line storage organization data change, the version number is updated.
Specifically, the query engine further comprises:
Rule defines memory module, for defining and storing the resolution rules of the cascade querying condition；
Querying condition parsing module is cascaded, for receiving data the cascade querying condition of consumer entering, according to the rule
The resolution rules for then defining the cascade querying condition that memory module defines, parse the cascade querying condition；
Enquiry module, according to the cascade querying condition that the cascade querying condition parsing module parsing obtains, step by step to institute
The columns group for stating the storage of column data memory module is inquired, and the column data position mark for meeting the cascade querying condition is obtained
Know.
Specifically, the cache unit further comprises:
Data memory module is inquired, joins querying condition, the targeted tables of data version of the cascade querying condition for storage level
The column data station location marker that this information and inquiry obtain;
Consistency check module, the grade joint investigation of cascade querying condition and the storage for inspection data consumer entering
Whether inquiry condition consistent and the tables of data version information of inspection data customer demand inquiry and the tables of data version of storage are believed
It whether consistent ceases；
Query result sending module, for when the judging result of consistency judgment models be it is consistent when, by the inquiry number
The row reading unit is sent to according to the column data station location marker corresponding with the cascade querying condition that memory module stores.
More specifically, the consistency check module further comprises:
Querying condition inspection module is cascaded, if in and cascade querying condition identical according to cascade querying condition character string
Not comprising non-idempotence function, then consistent rule is ask in decision level joint investigation, and item is inquired in the cascade of inspection data consumer entering
Whether part is consistent with the cascade querying condition of the storage；
Target data inspection module, tables of data version information and the storage for the inquiry of inspection data customer demand
Tables of data version information it is whether consistent.
Invention additionally discloses a kind of column memory storage querying methods, which comprises
Each column data in creation columns group storage line storage organization in memory；
Column data in the line storage organization is loaded into the columns group of the creation；
According to the cascade querying condition that data consumer inputs, the column data being loaded into columns group is looked into
It askes, obtains the column data station location marker for meeting the cascade querying condition；
Line number and the column data are obtained in the line storage organization according to the column data station location marker of the acquisition
The consistent All Datarows information of station location marker, is sent to data consumer.
Preferably, the method also includes:
By cascade querying condition that one query uses, the columns group being directed to, column data station location marker deposit caching is obtained,
When the cascade querying condition of data consumer input is identical as the cascade querying condition of storage, and the cascade querying condition pair
When the columns group answered does not change, according to the column data station location marker for meeting the cascade querying condition of the storage, obtain
Take line number and the consistent All Datarows information of the column data station location marker；
Set memory stores the period, deletes the column data for exceeding the memory storage period in columns group；
Caching period is set, the data for exceeding the caching period in caching are deleted；
It sets read lock and writes lock, control the read-write locking and unlocking to the columns group.
Detailed, the method for each column data in the columns group of the creation in memory storage line storage organization is specific
Are as follows:
The column data of column storage of array different lines is created, with the field of column data respective column in line storage organization
The name nominating columns group, the column array index are column data station location marker；
Tables of data is created, the columns group for belonging to identical line storage organization is stored in identical tables of data；
Multiple tables of data can be created, each tables of data includes multiple columns groups, and each columns group saves
The data of a column, the line number phase in the column data station location marker and its corresponding line storage organization in line storage organization
Together；
Column data is preferentially stored in column data memory module using simple types.
Preferably, the version number information for storing each tables of data, when the column data in the tables of data changes
When, update the version number information.
It is detailed, the cascade querying condition inputted according to data consumer, to the column being loaded into columns group
Data are inquired, and the method for meeting the column data station location marker of the cascade querying condition is obtained specifically:
Define and store the resolution rules of the cascade querying condition；
The cascade querying condition for receiving data consumer input, according to the cascade querying condition resolution rules of the storage,
Parse the cascade querying condition；
According to the cascade querying condition after the parsing, the data stored in the columns group are inquired step by step, are obtained
Take the column data station location marker for meeting the cascade querying condition.
In more detail, the judgement is when the cascade querying condition of data consumer input and the cascade inquiry item of storage
Part is consistent, and the not changed method of the corresponding columns group of the cascade querying condition specifically:
Determine if according to non-idempotence function is not included in and cascade querying condition identical if cascade querying condition character string
The consistent rule of querying condition is cascaded, item is inquired in the cascade querying condition of inspection data consumer entering and the cascade of the storage
Whether part is consistent；
If data consumer requires the tables of data version information of inquiry consistent with the tables of data version information of the storage,
Determine that the corresponding columns group of the cascade querying condition does not change.
The present invention is stored in a manner of columns group, is looked by the way that the data in line storage organization are loaded into memory
Qualified data are ask, and obtain the column data station location marker (i.e. the subscript of array position where data) of the data, are passed through
The mark, which obtains, meets the line number of querying condition, and then obtains full line data, the advantage of doing so is that, it is not needed in inquiry time
After going through the information that every a line takes every a line, calculating is compared to target column, but directly against the column in row storage organization
All data put it into array and carry out unified comparison, obtain the line number for the condition that meets, then obtain the line number of the condition of satisfaction
According to, be greatly saved traversal and every row is carried out the time that value compares calculating, to improve the efficiency of inquiry；The present invention simultaneously
It provides cascade querying condition freely to edit for data consumer, gradually reduces query context by inquiring step by step, further increase
Search efficiency；Preferably, the present invention also provides cachings, for saving the status information inquired and result information in the past, when next
When secondary connection querying condition arrives, if querying condition, target column array do not change, directly according to the inquiry of storage
Columns group station location marker reads row information and is sent to data consumer, search efficiency is further improved, in conclusion of the invention
Provide a kind of column memory storage inquiry mode for supporting high-speed and high-efficiency to inquire.
Detailed description of the invention
Fig. 1 is a kind of structural schematic diagram of the column memory storage inquiry unit of the embodiment of the present invention one；
Fig. 2 is the structural schematic diagram of two devices of the embodiment of the present invention；
Fig. 3 is a kind of column memory storage inquiry unit workflow schematic diagram that the embodiment of the present invention three combines example；
Fig. 4 is a kind of method flow diagram of the column memory storage querying method of the embodiment of the present invention four；
Fig. 5 is the method flow diagram of five method of the embodiment of the present invention；
Fig. 6 is that how resolution stage joins querying condition and carries out the method flow diagram of consistency judgement the embodiment of the present invention six；
Fig. 7 is that cascade query hierarchy schematic diagram is carried out in the embodiment of the present invention six.
Specific embodiment
Carry out the embodiment that the present invention will be described in detail below in conjunction with schema and embodiment, how the present invention is applied whereby
Technological means solves technical problem and reaches the realization process of technical effect to fully understand and implement.
As shown in Figure 1, the embodiment of the present invention one discloses a kind of column memory storage inquiry unit, including with flowering structure:
Memory storage unit 1, for creating each column data in columns group storage line storage organization in memory.
Since memory source space is limited, using each column data of simple storage of array.Structure of arrays only has subscript and value,
Since its structure is simple, so occupying less memory source.
Data load units 2, for the target column data in the line storage organization to be loaded into the memory storage
Unit 1.
Data load units obtain from data source there and load the target column data inquired.The week of loading
The size that phase loads has data source to determine, data source can issue it to data consumer and be mounted with which data, data consumption
Person can inquire the data of loading according to the information of publication.Data load units are carried out according to the load command of data source
Data load.
General line storage organization is as shown in table 1 below, and one has student's table of 10,000 rows.
1 line storage organization example of table
| Line number | Name (name) | Age (age) | Class (class) |
| 0 | Zhang | 17 | 1 |
| 1 | Lee | 22 | 2 |
| 2 | Mr. Wang | 21 | 3 |
| 3 | Zhao | 19 | 2 |
| … | … | … | … |
| 9999 | Qian | 21 | 1 |
To upper table 1, all rows are found out according to age=29, by the way of row storage are as follows:
(1) the row pointer to object that every a line gets every a line is traversed
(2) value that age column are got by row pointer to object may be according to the index of column age if it is universal architecture is used
Or title is indexed value.
(3) due to column generally according to universal architecture store, need to each age cell carry out (int) type conversion or
Compare lookup, the int and 29 after conversion carries out comparison of equalization, judges whether it is selected row, line pointer is put into knot in this way
In fruit queue.
(4) search is completed, and obtains whole result line pointers.
Using each column data in the column storage of array line storage organization, i.e., by name (name) conduct in upper table 1
One columns group, age (age) are used as a columns group, and class (class) is used as a columns group.
Name, age, class tri- is arranged as name column array, age columns group and class and is arranged by data load units
Array is loaded into memory storage unit.
When data load units are loaded with the data in multiple line storage organizations into memory storage unit, can introduce
Tables of data DataTable concept, the creation columns group set of memory storage unit can be described as tables of data DataTable, and one
DataTable includes multiple columns groups, and the column data of a column in line storage organization is saved in a columns group, and data load single
Respectively equipping in one line storage organization is loaded into each columns group in one DataTable by member.
It can be with the name nominating tables of data of line storage organization, with the name nominating columns group of column when loading.
Query engine 3, the cascade querying condition for being inputted according to data consumer fill the data load units 2
The column data for being loaded into the memory storage unit 1 is inquired, and the column data position mark for meeting the cascade querying condition is obtained
Know, is sent to row reading unit.
Cascade querying condition is that data consumer is write according to the data query requirements of itself, there is querying condition group one by one
At query engine can directly calculate the column data progress in columns group according to these querying conditions, avoid capable storage
First traversal is gone in structure, then takes the target column in row, then calculate, and meets the troublesome calculation process of record line number after design conditions.
After query engine is completed according to querying condition to the calculating of all column datas in columns group, by the columns for the condition that meets
It is recorded according to station location marker, is sent to row reading unit.
Row reading unit 4, the column data station location marker obtained according to the query engine 3 is in the line storage organization
Line number and the consistent All Datarows information of the column data station location marker are obtained, data consumer is sent to.
After obtaining column data station location marker, need to know according to column data station location marker corresponding row storage knot
Then line number in structure is expert at by line number and obtains all field informations of the row row in storage organization, is sent to data consumer.
Preferably, described in order to improve the search efficiency for being directed to same stages connection querying condition and being directed to same number of columns evidence
Device further include:
Cache unit 5, for storing the inquiry data of the query engine 3, when the cascade of data consumer input
Querying condition is consistent with the cascade querying condition of storage, and when the corresponding columns group of the cascade querying condition does not change, will
The column data station location marker for meeting the cascade querying condition of the storage is sent to row reading unit.
Due in actual operation, it may appear that data consumer carries out the inquiry of the same terms for same columns group,
In order to improve search efficiency, repeated work is avoided, cache unit 5 of the invention is responsible for saving each inquiry data, described
Inquiry data mainly include the content of three aspects, a querying condition inquired for this；Two inquire the columns group being directed to for this
Attribute information, three be to use this querying condition column data station location marker obtained to columns group.When the first two information is
When changing, then directly acquire column data station location marker be sent to row reading unit go obtain corresponding line data, be sent to number
According to consumer, the consistency checking efficiency of two kinds of information to be exceeded, each data in columns group are calculated again, therefore is slow
Memory cell is conducive to improve the working efficiency of apparatus of the present invention.
Preferably, in order to effectively use and control memory source, guarantee that system operates normally, described device further include:
Period control unit 6 deletes for storing the period for the memory storage unit set memory and exceeds institute in array
The column data for stating the memory storage period is also used to set caching period for the cache unit, deletes and exceed institute in cache unit
State the data cached of caching period.
Since data are constantly loaded into memory storage unit, finally when facing huge data volume by memory resource limitation
Vast resources can be consumed, efficiency is caused to decline, and some data are just of little use after inquiring one period, it can be with
It is cleared up in memory eases up and deposits.Therefore set memory storage period and caching period, the data for exceeding the period is deleted, guarantee this hair
The normal operation of bright device.
Preferably, in order to guarantee that the correctness when being written and read to memory storage unit 1, described device are also wrapped
It includes:
Control unit 7 is locked, for setting read lock to the memory storage unit 1 and writing lock, controls the memory storage
The read-write locking and unlocking of unit 1.
Read lock is the correctness in order to ensure read-write operation with lock is write, and the mode of locking may be set according to actual conditions,
Can such as chains be carried out to the tables of data in memory storage unit, the columns group in memory storage unit can also be locked.Locking
The locking principle of control unit is to guarantee that the same moment can only have a load thread to be in active state, and load thread is not tied
Others load thread and reading thread are all in wait state when beam；Any number of inquiry operations are supported concurrently to execute, but
When inquiry is not finished, load thread is waited for.
Based on the embodiment of the present invention one, for the working principle of further description apparatus of the present invention, spy provides this
The embodiment two of invention, as shown in Figure 2.
Memory storage unit 1 further comprises:
Column data memory module 11, for creating the column data of column storage of array different lines, with the column data in line
The field name of respective column names the columns group in storage organization, and the array index is column data station location marker.
The memory storage unit may include multiple column data memory modules, and each column data memory module saves row
The data of a column, the line number phase in the column data station location marker and its corresponding line storage organization in formula storage organization
Together.
Column data station location marker is the array index of the column data position, and array index starts with 0, according to array
Subscript takes the line number in its corresponding line storage organization, if line number is also to start with 0, takes identical line number, if
Line number is started with 1, then takes the line number of " subscript+1 ".
Column data is preferentially stored in column data memory module using simple types.
There are many selections, such as int type data storing by column, int [] or Integer [] can be used,
The common storage organizations such as List<Integer>.
The int [] that initialization is one 10,000,000 only needs 15ms, and needs 400ms, difference more than 20 using Integer []
Times, it is often more important that committed memory group member, int [] only needs 49MB, and Integer [] needs 228MB, such as uses
Then internal structure is equal to Object [] ArrayList<Integer>, and system and storage can only be greater than Integer [], knot
It is more effective storage organization by being self-evident, therefore using simple types storage array, common simple types is such as
Int, bool, double.
Tables of data memory module 12 is identical belonging to of storing by the column data memory module 11 for creating tables of data
The columns group of line storage organization is stored in identical tables of data.
The memory storage unit may include multiple tables of data memory modules, and each tables of data memory module may include multiple
Column data memory module.
Version number is arranged in tables of data to save in each tables of data memory module 12, when the tables of data is corresponding
When line storage organization data change, the version number is updated.
When needing to import multiple line storage tables in memory, column in each line storage table of expression for clarity,
Tables of data is preferably used, columns group and tables of data are subordinate relation, and a tables of data may include multiple columns groups, a line
Storage table corresponds to a tables of data, the data comprising all column in line storage in tables of data, the column data storage of each column
For a columns group.
Data load units 2, for the target column data in the line storage organization to be loaded into the memory storage
Columns group in unit 1.
Query engine 3 further comprises:
Rule defines memory module 31, for defining and storing the resolution rules of the cascade querying condition.
The cascade querying condition resolution rules of definition, i.e., pair of operator and general-purpose operation symbol in definition cascade querying condition
It should be related to.
It defines cascaded operational symbol corresponding relationship and provides convenience when can join querying condition for data consumer input stage, it is described
Corresponding relationship be it is customized, can change according to the actual situation cascaded operational symbol title.
Corresponding relationship example as shown in Table 2 is provided in the present invention.
2 cascaded operational of table symbol accords with corresponding relationship with general-purpose operation
Using the corresponding relationship recorded in table 2, cascade querying condition is write, such as shown below:
Eq (' class', 2) .bw (' age', 18,20) .eq (' sex', 0) .like (' address ', ' Haidian % % ')
Querying condition parsing module 32 is cascaded, for receiving data the cascade querying condition of consumer entering, according to described
Rule defines the resolution rules for the cascade querying condition that memory module 31 defines, and parses the cascade querying condition.
According to the corresponding relationship in table 2, above-mentioned cascade querying condition is parsed.
The meaning of above-mentioned cascade querying condition statement are as follows: class 2, age 18 (>=) between 20 (<20), gender be
It include the record in Haidian in male's (0 is male), address.
Enquiry module 33, according to the cascade querying condition that cascade querying condition parsing module 32 parsing obtains, step by step
The columns group of the column data memory module 11 storage is inquired, the column data position for meeting the cascade querying condition is obtained
Set mark.
This cascade querying condition is from left to right executed, every primary result of execution just becomes smaller once, has arrived most time-consuming
When like is operated, record strip number has greatly reduced, and the number for executing like is also also considerably reduced, and such overall efficiency is just
It is many to have promotion.
Row reading unit 4, the column data station location marker obtained according to the query engine 3 is in the line storage organization
Line number and the consistent All Datarows information of the column data station location marker are obtained, data consumer is sent to.
Cache unit 5 further comprises:
Data memory module 51 is inquired, joins querying condition, the targeted tables of data of the cascade querying condition for storage level
The column data station location marker that version information and inquiry obtain.
In order to avoid carrying out repeating inquiry to the same terms same queries target data, by inquiring data memory module,
The column data subscript that the querying condition inquired every time, inquiry target data, inquiry obtain is stored, so as to data consumer
When being inquired for same queries condition and target data, the column data subscript for obtaining the condition that meets is directly transmitted, improves effect
Rate.
Consistency check module 52, the cascade of cascade querying condition and the storage for inspection data consumer entering
Whether querying condition consistent and the tables of data version of the tables of data version information of inspection data customer demand inquiry and storage
Whether information is consistent.
Consistency check module is exactly the querying condition that whether there is data consumer input in caching for checking, and
Whether inquiry target data changes, under the column data for meeting querying condition by that can directly transmit storage after examining
Mark.Therefore consistency check module 52 should carry out the inspection of two aspects, further comprise:
Querying condition inspection module 521 is cascaded, if for and cascade inquiry item identical according to cascade querying condition character string
Not including non-idempotence function in part, then consistent rule, the grade joint investigation of inspection data consumer entering are ask in decision level joint investigation
Whether inquiry condition is consistent with the cascade querying condition of the storage.
Judge whether the cascade querying condition of data consumer input is identical as the cascade querying condition of the storage, passes through
Two condition criterions, whether one is identical for character string, and whether two be comprising non-idempotence function.Character string is identical successively to be compared
Character string.Non- idempotence function is to repeat the different function of implementing result, although its character string is identical, due to non-idempotence letter
Several characteristics, the result that causes it to repeat occur it is different, therefore include the querying condition of non-idempotence function in fact
Changed, it is not identical for should surveying.Such as now () function is to obtain current time, it being performed a plurality of times the result is that
Inconsistent, it is meant that if including now () function in query expression, such query result cannot be reused, be needed
Query execution is carried out every time.
When treating non-idempotence function, there are also a kind of processing modes, i.e., inquiry data memory module 51 is to inquiry item
Whether judge in querying condition when part is stored comprising non-idempotence function, if comprising, not to the querying condition and
Relevant information corresponding with the querying condition is stored.
Target data inspection module 522, for inspection data customer demand inquiry tables of data version information with it is described
Whether the tables of data version information of storage is consistent.
Querying condition is consistent, but target data changes, and also results in query result and changes, it is therefore desirable to mesh
Mark data, which carry out consistency check, will be updated since tables of data stores version information when data change in tables of data
Version information, therefore only need simple comparison data table version information that consistency check can be completed.
Query result sending module 53, for being looked into described when the judging result of consistency judgment module 52 is consistent
It askes the column data station location marker corresponding with the cascade querying condition that data memory module 51 stores and is sent to the row reading list
Member.
Period control unit 6 deletes for storing the period for the memory storage unit set memory and exceeds institute in array
The column data for stating the memory storage period is also used to set caching period for the cache unit, deletes and exceed institute in cache unit
State the data cached of caching period.
Control unit 7 is locked, for setting read lock to the memory storage unit 1 and writing lock, controls the memory storage
The read-write locking and unlocking of unit 1.
To better illustrate apparatus of the present invention and working principle, for the table 1 in the present invention, based on embodiment one, implement
Example two provides the embodiment of the present invention three, is described in detail, as shown in Figure 3.
Step S101: column data memory module 11 creates columns group according to line storage organization.
Including name column array, age columns group, class columns group.Each columns group is stored with simple data type respectively
In data.As shown in table 3.
3 column storage of array schematic diagram of table
Step S102: tables of data memory module 12 creates tables of data, saves each columns group created in step S101, version
Number be studentV1.0.
Step S103: data consumer defines the rule that memory module 31 defines according to rule, writes cascade querying condition.
The cascade querying condition write is " eq (' class', 2) .bw (' age', 18,20) "
Step S104: cascade querying condition parsing module 32 defines the rule that memory module 31 defines, docking according to rule
Even querying condition is parsed.
The result of parsing is 2 to inquire class, data of the age between 18 to 20.
Step S105: the querying condition that enquiry module 33 is parsed according to querying condition parsing module 32 stores tables of data
Class columns group, the age columns group stored in module 12 is inquired step by step.
Class columns group is first inquired ,=2 data subscript is obtained, is 1,3 ... ...；
Age columns group is inquired on this basis, obtains the array index between 18-20, is 3 ... ....
Step S106: row reading unit 4 reads the institute of the row according to the column data subscript of acquisition in line storage organization
There is row data information to be sent to data consumer.
Row data-reading unit 4 read line number be 3 ... row data information, be sent to data consumer.
Step S107: inquiry data memory module 51 stores the querying condition and the corresponding mesh of querying condition of this inquiry
Mark data version information and column data array index.
The querying condition of storage are as follows: " eq (' class', 2) .bw (' age', 18,20) "；
The target data version information of storage are as follows: student1.0；
The column data subscript of storage are as follows: 3 ....
Step S108: consistency check module 52 receives new cascade querying condition, carries out to new cascade querying condition
Consistency check enters step S109 by examining, not verified to enter step S104.
Step S109: column data subscript corresponding with the concatenation condition is obtained in inquiry data memory module 51, is entered
Step S106.
The embodiment of the present invention four discloses a kind of column memory storage querying method, as shown in Figure 4, which comprises
Step S201: each column data in columns group storage line storage organization is created in memory.
It should be noted that is created herein is empty columns group, the purpose is to store the lattice of line storage organization column
Data.
Step S202: the column data in the line storage organization is loaded into the columns group of the creation.
Step S203: the cascade querying condition inputted according to data consumer, to the columns being loaded into columns group
According to being inquired, the column data station location marker for meeting the cascade querying condition is obtained.
Define and store the resolution rules of the cascade querying condition；Item is inquired in the cascade for receiving data consumer input
Part parses the cascade querying condition according to the cascade querying condition resolution rules of the storage；According to the grade after the parsing
Join querying condition, the data stored in the columns group are inquired step by step, obtains the column for meeting the cascade querying condition
Data Position mark.
Step S204: line number and institute are obtained in the line storage organization according to the column data station location marker of the acquisition
The consistent All Datarows information of column data station location marker is stated, data consumer is sent to.
Due to when carrying out column storage of array, being the row sequential storage according to each column data in line storage organization, because
This column data station location marker (array index i.e. where the column data) and its line number for being located at row in line storage organization
It is corresponding, it should be noted that array index since 0, if the line number of line storage organization also since 0, column data
The array index at place as its line number, if the line number of line storage organization since 1, by the number where column data
Group subscript is used as its line number after adding 1.
Obtain the line number where column data, so that it may which the information of the full line taken in line storage organization according to line number is sent out
Data consumer is given, the search efficiency of data consumer is improved.
In order to further increase same queries condition, for the search efficiency of same target columns group, guarantee rationally utilization
With control memory source, guarantee the read-write correctness of columns group, it is preferred that can by cascade querying condition that one query uses,
For columns group, obtain column data station location marker deposit caching, when data consumer input the cascade querying condition with
The cascade querying condition of storage is identical, and when the corresponding columns group of the cascade querying condition does not change, according to the storage
The column data station location marker for meeting the cascade querying condition obtain line number and the column data station location marker is consistent all
Row data information.
Set memory stores the period, deletes the column data for exceeding the memory storage period in columns group.
Caching period is set, the data for exceeding the caching period in caching are deleted.
It sets read lock and writes lock, control the read-write locking and unlocking to the columns group.
In the following, the embodiment of the present invention five is provided, as shown in figure 5, described method includes following steps:
Step S301: each column data in columns group storage line storage organization is created in memory.
The column data of column storage of array different lines is created, with the field of column data respective column in line storage organization
The name nominating columns group, the column array index are column data station location marker.
Column data is preferentially stored in column data memory module using simple types.
Column data is stored using simple types, it is possible to reduce the memory source that data storage occupies, arithmetic speed also have
Significantly promoted.
Step S302: the columns group for belonging to identical line storage organization is stored in identical data by creation tables of data
Table.
Multiple tables of data can be created, each tables of data includes multiple columns groups, and each columns group saves
The data of a column, the line number phase in the column data station location marker and its corresponding line storage organization in line storage organization
Together.
The version number information for storing each tables of data is updated when the column data in the tables of data changes
The version number information.
Step S303: the columns column data in the line storage organization being loaded into the tables of data of the creation
Group.
Column data in identical line storage organization is loaded into a tables of data, may include multiple columns in the tables of data
Group, for loading each column data in the line storage organization.The corresponding columns group of every column data.
Step S304: the cascade querying condition inputted according to data consumer judges in caching with the presence or absence of same
Cascade querying condition, then enter step S305 if it does not exist, then enter step S306 if it exists.
The previous data consumer within caching period is stored in caching to look into target column data according to querying condition
The relevant information of inquiry, if finding, the querying condition of this inquiry was once used, further judged that the target column data being directed to is
It is no identical；If the querying condition of this inquiry never occurs, inquired according to new querying condition.
Step S305: according to the cascade querying condition of the consumer entering, to the data being loaded into columns group
It is inquired, obtains the column data station location marker for meeting the cascade querying condition, the inquiry data of this inquiry are stored in slow
It deposits, enters step S308.
This inquiry is new inquiry, then in this inquiry of completion, data deposit caching will be inquired, when appearance and this
When inquiring identical inquiry, the column data mark that inquiry obtains can be directly taken out, into the reading of line data, can be improved effect
Rate.
Inquiry data of this inquiry include the cascade querying condition used, this tables of data version for being directed to of inquiry
The column data station location marker that information and this inquiry obtain.
Step S306: judging whether tables of data version information corresponding with the cascade querying condition changes in caching,
S305 is entered step if changing, if entering step S307 there is no variation.
If tables of data version information is changed, illustrates that target column data may have occurred update, be loaded again
Or other variations have occurred, therefore inquiry is changed, it cannot query result before use, it is therefore desirable to return step
S305。
Step S307: column data station location marker corresponding with the cascade querying condition is obtained in the buffer.
Step S308: line number and institute are obtained in the line storage organization according to the column data station location marker of the acquisition
The consistent All Datarows information of column data station location marker is stated, data consumer is sent to.
If data consumer requires the tables of data version information of inquiry consistent with the tables of data version information of the storage,
Determine that the corresponding columns group of the cascade querying condition does not change.
It illustrates how to complete rule definition and parsing to cascade querying condition in order to clearer, and how to determine to inquire
Whether condition is consistent, provides the embodiment of the present invention six, as shown in Figure 6.
It is often many condition in one query, such as " class is 2 classes, and the age, gender was male between 18~20 for lookup
Property student ", inquiry parsing need one by one condition execute, in the case where any field is fair without index, such processing
For program it is that very easily, just condition executes one by one for it, it is apparent that these three conditions are the relationships of and, the 1st
Condition executes completion, should not should do in the 1st result on all rows in the 2nd condition of execution, using grade
Connection expression formula carrys out descriptive level connection querying condition, it can be achieved that inquiry step by step to each condition, further increases search efficiency.Therefore this
In set cascaded expression writing mode.
Step S401: defining and stores the resolution rules of the cascade querying condition.
The cascade querying condition resolution rules of definition, i.e., pair of operator and general-purpose operation symbol in definition cascade querying condition
It should be related to.As shown in Table 2 above.
It defines cascaded operational symbol corresponding relationship and provides convenience when can join querying condition for data consumer input stage, it is described
Corresponding relationship be it is customized, can change according to the actual situation cascaded operational symbol title.
Step S402: the cascade querying condition of data consumer input is received, according to the cascade querying condition of the storage
Resolution rules parse the cascade querying condition.
It is placed on front according to simple querying condition, the principle that complicated querying condition is put behind writes cascade inquiry item
Part.
Cascade querying condition is from left to right executed, every primary result of execution just becomes smaller once, has arrived most time-consuming operation
When, record strip number has greatly reduced, and the number for executing like is also also considerably reduced, and such overall efficiency just has promotion
Very much, as shown in following Fig. 7.
It, can be according to by not time-consuming operation (such as equal) when user's input stage joins querying condition according to above-mentioned thinking
Preposition, time-consuming operation (like) postposition, whole efficiency just have promotion by a larger margin.
Step S403: according to the cascade querying condition after the parsing, judge the cascade querying condition character string and delay
Whether the character string for depositing middle storage is identical, then enters step S404 if they are the same, if not identical, enters step S405.
Judge whether character string is identical, it is only necessary to which verification can be obtained as a result, repeating no more one by one.
Step S404: judge to walk whether comprising non-idempotence function if entered comprising if in the cascade querying condition
Rapid S405 enters step S406 if not including.
In the case where character string is identical, if including non-idempotence function in character string, indicate that the character string is retouched
The cascade querying condition stated has been likely occurred variation.
Such as non-idempotence function now(), expression takes current state, and the query result data saved in caching, before being
The result data inquired not is current result data, therefore cannot directly acquire.
Step S405: determine that the cascade querying condition of data consumer input and the cascade querying condition in caching are different
It causes.
Step S406: determine that the cascade querying condition of data consumer input is consistent with the cascade querying condition in caching.
The detailed implementation of each step in this method, reference can be made to device part description above, therefore repeat no more.
Although disclosed herein embodiment it is as above, the content is not of the invention directly to limit
Protection scope.Any the technical staff in the technical field of the invention, do not depart from disclosed herein spirit and scope
Under the premise of, a little change can be made in the formal and details of implementation.Protection scope of the present invention, still must be with appended power
Subject to the range that sharp claim is defined.
Claims (13)
1. a kind of column memory storage inquiry unit, which is characterized in that described device includes:
Memory storage unit, for creating each column data in columns group storage line storage organization in memory；
Data load units, for the target column data in the line storage organization to be loaded into the memory storage unit；
Query engine, the cascade querying condition for being inputted according to data consumer are loaded into institute to the data load units
The column data for stating memory storage unit is inquired, and is obtained the column data station location marker for meeting the cascade querying condition, is sent
Give row reading unit；
Row reading unit obtains row according to the column data station location marker that the query engine obtains in the line storage organization
Number with the consistent All Datarows information of the column data station location marker, be sent to data consumer；
Described device further include:
Cache unit, for storing the inquiry data of the query engine, when item is inquired in the cascade of data consumer input
Part is consistent with the cascade querying condition of storage, and when the corresponding columns group of the cascade querying condition does not change, deposits described
The column data station location marker for meeting the cascade querying condition of storage is sent to row reading unit.
2. the apparatus according to claim 1, which is characterized in that described device further include:
Period control unit is deleted in array and is exceeded in described for storing the period for the memory storage unit set memory
The column data for storing the period is also used to set caching period for the cache unit, deletes in cache unit beyond described slow
Cycle of deposit it is data cached；
Control unit is locked, for setting read lock to the memory storage unit and writing lock, controls the memory storage unit
Read and write locking and unlocking.
3. the apparatus of claim 2, which is characterized in that the memory storage unit further comprises:
Column data memory module is stored in line with the columns group and is tied for creating the column data of column storage of array different lines
The field name of respective column names the columns group in structure, and the array index is column data station location marker；
Tables of data memory module stores the identical line that belongs to of column data memory module storage for creating tables of data
The columns group of structure is stored in identical tables of data；
The memory storage unit may include multiple tables of data memory modules, and each tables of data memory module may include multiple columns
According to memory module, each column data memory module saves the data of a column in line storage organization, the column data position
It identifies identical with the line number in its corresponding line storage organization；
The column data is preferentially stored in column data memory module using simple types.
4. device according to claim 3, it is characterised in that:
Version number is arranged in tables of data to save in each tables of data memory module, when the corresponding row of the tables of data
When formula storage organization data change, the version number is updated.
5. device according to claim 4, which is characterized in that the query engine further comprises:
Rule defines memory module, for defining and storing the resolution rules of the cascade querying condition；
Querying condition parsing module is cascaded, for receiving data the cascade querying condition of consumer entering, it is fixed according to the rule
The resolution rules for the cascade querying condition that adopted memory module defines, parse the cascade querying condition；
Enquiry module, according to the cascade querying condition that the cascade querying condition parsing module parsing obtains, step by step to the column
The columns group of data memory module storage is inquired, and the column data station location marker for meeting the cascade querying condition is obtained.
6. device according to claim 5, which is characterized in that the cache unit further comprises:
Data memory module is inquired, joins querying condition, the targeted tables of data version letter of the cascade querying condition for storage level
The column data station location marker that breath and inquiry obtain；
Consistency check module inquires item for the cascade querying condition of inspection data consumer entering and the cascade of the storage
Whether part consistent and the tables of data version information of inspection data customer demand inquiry and the tables of data version information of storage are
It is no consistent；
Query result sending module, for when the inspection result of consistency judgment models is consistent, the inquiry data to be deposited
The column data station location marker corresponding with the cascade querying condition of storage module storage is sent to the row reading unit.
7. device according to claim 6, which is characterized in that the consistency check module further comprises:
Querying condition inspection module is cascaded, if for not wrapping in and cascade querying condition identical according to cascade querying condition character string
Containing non-idempotence function, then consistent rule is ask in decision level joint investigation, the cascade querying condition of inspection data consumer entering with
Whether the cascade querying condition of the storage is consistent；
Target data inspection module, for the tables of data version information of inspection data customer demand inquiry and the number of the storage
It is whether consistent according to table version information.
8. a kind of column memory storage querying method, which is characterized in that the described method includes:
Each column data in creation columns group storage line storage organization in memory；
Column data in the line storage organization is loaded into the columns group of the creation；
According to the cascade querying condition that data consumer inputs, the column data being loaded into columns group is inquired, is obtained
Take the column data station location marker for meeting the cascade querying condition；
Line number and the column data position are obtained in the line storage organization according to the column data station location marker of the acquisition
Consistent All Datarows information is identified, data consumer is sent to；
The method also includes:
By cascade querying condition that one query uses, the columns group being directed to, column data station location marker deposit caching is obtained, works as number
It is identical as the cascade querying condition of storage according to the cascade querying condition of consumer entering, and the cascade querying condition is corresponding
When columns group does not change, according to the column data station location marker for meeting the cascade querying condition of the storage, row is obtained
Number with the consistent All Datarows information of the column data station location marker.
9. according to the method described in claim 8, it is characterized in that, the method also includes:
Set memory stores the period, deletes the column data for exceeding the memory storage period in columns group；
Caching period is set, the data for exceeding the caching period in caching are deleted；
It sets read lock and writes lock, control the read-write locking and unlocking to the columns group.
10. according to the method described in claim 9, it is characterized in that, the columns group of the creation in memory storage line storage
The method of each column data in structure specifically:
The column data of column storage of array different lines is created, with the field name of column data respective column in line storage organization
The columns group is named, the column array index is column data station location marker；
Tables of data is created, the columns group for belonging to identical line storage organization is stored in identical tables of data；
Multiple tables of data can be created, each tables of data includes multiple columns groups, and each columns group saves line
The column data of a column, the line number phase in the column data station location marker and its corresponding line storage organization in storage organization
Together；
Column data is preferentially stored in column data memory module using simple types.
11. according to the method described in claim 10, it is characterized by:
The version number information for storing each tables of data, when the column data in the tables of data changes, described in update
Version number information.
12. according to the method for claim 11, which is characterized in that item is inquired in the cascade according to data consumer input
Part inquires the column data being loaded into columns group, obtains the column data position for meeting the cascade querying condition
The method of mark specifically:
Define and store the resolution rules of the cascade querying condition；
The cascade querying condition for receiving data consumer input, according to the cascade querying condition resolution rules of the storage, parsing
The cascade querying condition；
According to the cascade querying condition after the parsing, the data stored in the columns group are inquired step by step, are obtained full
The column data station location marker of the foot cascade querying condition.
13. according to the method for claim 12, which is characterized in that the cascade of data consumer input is worked as in the judgement
Querying condition is consistent with the cascade querying condition of storage, and the not changed method of the corresponding columns group of the cascade querying condition
Specifically:
Determine to cascade if according to non-idempotence function is not included in and cascade querying condition identical if cascade querying condition character string
The consistent rule of querying condition, the cascade querying condition of inspection data consumer entering and the cascade querying condition of the storage are
It is no consistent；
If data consumer requires the tables of data version information of inquiry consistent with the tables of data version information of the storage, determine
The corresponding columns group of the cascade querying condition does not change.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310744231.9A CN104750727B (en) | 2013-12-30 | 2013-12-30 | A kind of column memory storage inquiry unit and column memory storage querying method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310744231.9A CN104750727B (en) | 2013-12-30 | 2013-12-30 | A kind of column memory storage inquiry unit and column memory storage querying method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104750727A CN104750727A (en) | 2015-07-01 |
| CN104750727B true CN104750727B (en) | 2019-03-26 |
Family
ID=53590426
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310744231.9A Active CN104750727B (en) | 2013-12-30 | 2013-12-30 | A kind of column memory storage inquiry unit and column memory storage querying method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104750727B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105589969A (en) * | 2015-12-23 | 2016-05-18 | 浙江大华技术股份有限公司 | Data processing method and device |
| CN107092624B (en) * | 2016-12-28 | 2022-08-30 | 北京星选科技有限公司 | Data storage method, device and system |
| CN107436767A (en) * | 2017-07-31 | 2017-12-05 | 杭州安恒信息技术有限公司 | The optimization method that idempotent operates in a kind of asynchronous framework |
| CN110069487A (en) * | 2017-09-28 | 2019-07-30 | 北京国双科技有限公司 | A kind of data processing method, apparatus and system |
| CN109101516B (en) | 2017-11-30 | 2019-09-17 | 新华三大数据技术有限公司 | A kind of data query method and server |
| CN109165378A (en) * | 2018-08-15 | 2019-01-08 | 北京天安智慧信息技术有限公司 | Sophisticated functions Report Customization method and system |
| JP7199522B2 (en) * | 2018-10-09 | 2023-01-05 | タブロー ソフトウェア，インコーポレイテッド | Correlated incremental loading of multiple datasets for interactive data prep applications |
| CN109445945B (en) * | 2018-10-29 | 2023-09-19 | 努比亚技术有限公司 | Memory allocation method of application program, mobile terminal, server and storage medium |
| US11386089B2 (en) | 2020-01-13 | 2022-07-12 | The Toronto-Dominion Bank | Scan optimization of column oriented storage |
| CN117931802B (en) * | 2024-01-25 | 2024-09-13 | 南京雀翼信息科技有限公司 | High-speed writing and reading system and method for multiple data sources |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102136005A (en) * | 2011-03-29 | 2011-07-27 | 北京航空航天大学 | Data searching method and device |
| CN103246498A (en) * | 2013-05-13 | 2013-08-14 | 浪潮集团山东通用软件有限公司 | Memory storage structures supporting relational data parallel processing and achieving method thereof |
-
2013
- 2013-12-30 CN CN201310744231.9A patent/CN104750727B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102136005A (en) * | 2011-03-29 | 2011-07-27 | 北京航空航天大学 | Data searching method and device |
| CN103246498A (en) * | 2013-05-13 | 2013-08-14 | 浪潮集团山东通用软件有限公司 | Memory storage structures supporting relational data parallel processing and achieving method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104750727A (en) | 2015-07-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104750727B (en) | A kind of column memory storage inquiry unit and column memory storage querying method | |
| US9875280B2 (en) | Efficient partitioned joins in a database with column-major layout | |
| US8935575B2 (en) | Test data generation | |
| US9129002B2 (en) | Dividing device, dividing method, and recording medium | |
| CN104361113B (en) | A kind of OLAP query optimization method under internal memory flash memory mixing memory module | |
| CN107943952B (en) | Method for realizing full-text retrieval based on Spark framework | |
| EP2354921A1 (en) | Hybrid evaluation of expressions in DBMS | |
| US20170255673A1 (en) | Batch Data Query Method and Apparatus | |
| CN108664516A (en) | Enquiring and optimizing method and relevant apparatus | |
| CN101901265B (en) | Objectification management system of virtual test data | |
| CN102236672A (en) | Method and device for importing data | |
| CN107038222A (en) | Database caches implementation method and its system | |
| CN101251825A (en) | Device and method for generating test use case | |
| CN103019691A (en) | Transformation method for extract, transform and load (ETL) operation relation graph and implementation system thereof | |
| JP2007334627A (en) | Service base software design support method, and device therefor | |
| EP3293644B1 (en) | Loading data for iterative evaluation through simd registers | |
| CN106844320A (en) | A kind of financial statement integration method and equipment | |
| JP2014502756A (en) | Apparatus and method for mass data storage based on tree structure | |
| JP5936135B2 (en) | Information processing apparatus, information processing method, and program | |
| CN100395752C (en) | Report data collection system and method | |
| CN101719162A (en) | Multi-version open geographic information service access method and system based on fragment pattern matching | |
| CN115062028B (en) | Method for multi-table join query in OLTP field | |
| CN111126619A (en) | Machine learning method and device | |
| CN107888686B (en) | User data validity verification method located at HBase client | |
| CN105630997A (en) | Data parallel processing method, device and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180705 Address after: 110179 room 519, 2-1 Gao Ge Road, Hunnan New District, Shenyang, Liaoning. Applicant after: Yiyang Computer Technology Co.,Ltd. Shenyang Address before: No. 1 building, hi tech Development Zone, Songshan Road, Nangang District, Harbin, Heilongjiang Applicant before: BOCO INTER-TELECOM Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230704 Address after: 150001 Building 1, High tech Development Zone, Songshan Road, Nangang District, Harbin City, Heilongjiang Province Patentee after: BOCO INTER-TELECOM Co.,Ltd. Address before: 110179 room 519, 2-1 Gao Ge Road, Hunnan New District, Shenyang, Liaoning. Patentee before: Yiyang Computer Technology Co.,Ltd. Shenyang |
|
| TR01 | Transfer of patent right |