CN100357952C - Binary data access method - Google Patents
Binary data access method Download PDFInfo
- Publication number
- CN100357952C CN100357952C CNB2005101243552A CN200510124355A CN100357952C CN 100357952 C CN100357952 C CN 100357952C CN B2005101243552 A CNB2005101243552 A CN B2005101243552A CN 200510124355 A CN200510124355 A CN 200510124355A CN 100357952 C CN100357952 C CN 100357952C
- Authority
- CN
- China
- Prior art keywords
- field
- data
- record
- index
- file
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention discloses a binary data access method for establishing record index files and field index files to each binary data file; the method also comprises the following steps: step A, the data type of a field to be accessed and the recorded data type are obtained; step B, a record index corresponding to the record matched with the recorded data type of the field to be accessed is found from a record index file; step C, the index of the field to be accessed is obtained from the field index file according to the record index found in the step B and the data type of the field to be accessed; step D, the field to be accessed is read from a binary data file according to the index of the field to be accessed, which is obtained in the step C. The technical project of the present invention can raise the analytic efficiency to a binary data file.
Description
Technical field
The present invention relates to the communication data treatment technology, particularly relate to a kind of access method of binary data.
Background technology
In communication system, often use binary format that information is encoded.(Data Supply, DS) after the generation, through storage or transmission, (Data Analysis DA) carries out analyzing and processing to binary file by data analysis module by data generating module.
In a binary file, generally include some records, every record comprises several fields again.Field can be various data types, causes record also to have various data types thus.The difference of data type has caused data shared memory length difference in binary file.Because data are storages continuously in binary file, in order to guarantee the DA correctly starting point and the terminating point of identification record and field, by the data type of this record of recording feature representation of regular length, represent the length of this record in place that the data of each record begin again by the record length sign indicating number of regular length.DA is indexing corresponding record data type in advance with in the concordance list that DS appoints according to condition code, further determines the data type of the field that comprised in the record according to the data type of record.
For example, in a data file, agreement recording feature code length is 1 byte, and record length is 1 byte; First field of first record of file is that length is the no symbol short field of 2 bytes, thisly just can determine that the field of length is called the static data types field according to data type; Second field is an adjustable length character type array field, the number of array element is to be determined by the value of first field, thisly need could determine that the field of length is called the dynamic data type field according to the data type of self and the value of other fields.If first writes down first field value is 5, the shared memory length of first record is 9 bytes so--condition code 1 byte, length code 1 byte, first field 2 bytes, second field 5 byte.DA begins to locate condition code in first byte according to data file, judges first field that the 3rd byte and the 4th byte are first records, judges second field that the 5th byte to the 9 bytes are first records in conjunction with the value of first field again; DA also begins to locate the length code of first field in second byte according to data file, judging from the 10th byte is the condition code of second record, and the 11st byte is the length code of second record.
In the prior art, in order to visit the specific fields in the record with specific data type, DA need write down one by one from the place that file begins and analyze, and each record that meets the type requirement is found out, up to handling whole binary data file.For example, suppose to have a kind of record of data type to comprise 5 fields, in a binary data file that 1000 records are arranged, belong to record 10 of this data type, be randomly dispersed in the binary data file.In order to visit these 10 records the 4th field separately respectively, need be from the beginning of binary data file, check successively whether the type codes of each record meets the data type of execution with the data type of judging current record, if meet then, find the field that to visit according to the field of data designated type analysis current record; If do not meet then calculate the address at the type codes place of next record according to the length code of current record; Then next record is repeated above inspection and judgement; So repeatedly up to complete binary data file of scanning.In the practical application, the number that writes down in each binary file reaches 10
4Even it is bigger; and often have and to carry out Conjoint Analysis to a plurality of binary data files of separate sources; if at every turn the analysis of binary data file is all needed to do once from the beginning to the end scanning; in inspection, judgement and the calculating in the time of will causing the big energy-handling capability of DA to be wasted in scanning, further can cause DA to analyze data capability and die down.
Summary of the invention
In view of this, fundamental purpose of the present invention is to provide a kind of access method of binary data file, each visit of binary file is all needed to scan from the file reference position avoiding, thereby improves the efficient of DA visit binary data file.
In order to achieve the above object, the invention provides a kind of access method of binary data, this method is set up recording indexes file and field index file for each binary data file, and this method also comprises:
Steps A is obtained the data type of field to be visited and the data type of place record;
Step B, the pairing recording indexes of record that from the recording indexes file, finds the data type with field to be visited place record to be complementary;
Step C obtains the index of field to be visited in the field index file according to the data type of recording indexes that is found among the step B and field to be visited;
Step D, the index of the field to be visited that obtains according to step C reads field to be visited from binary data file.
Wherein, the described pairing recording indexes of record that finds the data type with field to be visited place record to be complementary from the recording indexes file of step B is:
Step B1, the reading and recording index is as the current record index from the recording indexes file;
Step B2, determine the data type of the pairing record of current record index, and whether the data type of judging the pairing record of current record index meets the data type of field to be visited place record, if meet then think that the current record index is the pairing recording indexes of record that the data type with field to be visited place record is complementary, otherwise think that the current record index is not the pairing recording indexes of record that the data type with field to be visited place record is complementary;
Step B3 reads next bar recording indexes as the current record index from the recording indexes file, return execution in step B2.
Wherein, describedly set up the recording indexes file for binary data file and be:
The pairing recording indexes of record that will have different types of data leaves in the different recording indexes files, and with the sign of the pairing record data type of the stored recording indexes of each recording indexes file as this document;
Wherein, the described pairing recording indexes of record that finds the data type with field to be visited place record to be complementary from the recording indexes file of step B is:
Will with the recording indexes in the corresponding recording indexes file of data type of field to be visited place record, the pairing recording indexes of record that is complementary as data type with field to be visited place record
Wherein, step C comprises:
Step C1 is the address with the lead-in segment index address data items of current record index, and the data that read length and be a field index data length from the field index file are as current field index;
Step C2, determine the data type of the pairing field of current field index, and whether the data type of judging the pairing field of current field index meet the data type of field to be visited, if meet then think that current field index is the index of field to be visited, otherwise execution in step C3;
Step C3, the next field index that reads current record from the field index file is returned execution in step C2 as current field index.
Wherein, step D comprises:
Step D1 determines the data type of field to be visited according to the value of field data types data item in the field index of field to be visited, and determines the data length of field to be visited in binary data file;
Step D2 is the address with the value of field address data item in the field index of field to be visited, reads in the binary data file data that begin from appropriate address as field to be visited, and the data length that reads is the data length that obtains among the step D1.
Wherein, the address value of the record data categorical data item put data type in recording indexes, the data type of the pairing record of the described definite current record index of step B2 is:
From with the appropriate address reading of data of the corresponding dtd file of binary data file data type as record.
Wherein, the field data types data item in field index is deposited the address value of field data types, and step C2 is described to determine that the data type of the pairing field of current field index is:
From with the appropriate address reading of data of the corresponding dtd file of binary data file data type as field.
Wherein, step D1 is described determines that the data length of field to be visited in binary data file is:
Step D11 determines that according to field data types this field is static data types or dynamic data type, if static data types execution in step D12 then, otherwise execution in step D13;
Step D12 determines the data length of field to be visited in binary data file, execution in step D2 according to the definition of static data types;
Step D13 reads definite needed other information of the length of field to be visited in binary data file according to the dynamic data type definition, and then determines the length of field to be visited in binary data file.
Wherein, further comprise between step B2 and the step B3:
Judge whether to handle all recording indexes in the recording indexes file, if handle then stop this browsing process, otherwise execution in step B3.
Adopt the access method of binary data file provided by the present invention, only need carry out scanning whole file to set up concordance list in the analysis first time to file, just can locate fast the data in the binary data file later, thereby improve the efficient that DA analyzes binary file according to this concordance list.And, owing to a plurality of data files can be carried out index with same recording indexes file and field index file in actual applications, so DA can carry out Conjoint Analysis to the data file in a plurality of sources easily.
Description of drawings
Fig. 1 is the index structure of setting up in the access method of binary data file provided by the invention.
Fig. 2 is the process flow diagram of visit data in the access method of binary data file provided by the invention.
Embodiment
Core concept of the present invention is, when binary data file being carried out the processing first time, scan whole data file and set up the index of record and field location, thereby find required record and field fast by index when visiting this data file afterwards.
For making the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with drawings and the specific embodiments.
Please refer to Fig. 1, Fig. 1 is the index structure of setting up in the access method of binary data file provided by the invention.
The index that the method for fast access binary data file provided by the present invention need be set up is a two-stage index, is respectively field index and recording indexes.The corresponding field index of each field in the binary data file, the corresponding recording indexes of each record.All field index leave in the field index file according to the order of corresponding field in binary data file, and all recording indexes leave in the recording indexes file according to the order of corresponding record in binary data file.
Field index comprises two data item, is respectively field address and field data types.Wherein field address is the address of this field in binary data file, and field data types is a pointer, points to the storage space of depositing this field data types.For a record, the field data types of each field all is different.Recording indexes comprises three data item, is respectively lead-in segment index address, field number and record data type.Wherein the address of index in the field index file of first field of this record represented in lead-in segment index address, the field number is represented the field number that this record comprises, the record data type is a pointer, points to the storage space of depositing this record data type.
Above-described storage space typically refers to certain position in the dtd file.This file has comprised the definition of analyzing all data types that this binary data file will use, and it submitted to DA by DS before data analysis.
As DA for the first time during the Treatment Analysis binary data file, at first need binary data file is done once from the beginning to the end scanning to set up concordance list.
When handling first record, DA begins to locate the condition code of first record according to binary data file, determine the data type of first record, further revise corresponding to record data categorical data item in the recording indexes of first record and field number data item according to the data type of first record.DA also will be revised as the address of field index in the field index file of first field correspondence of first record corresponding to the lead-in segment index address data items in the recording indexes of first record.
According to the data type of first record of binary data file, DA can analyze the data type of each field in first record, and revises respectively in view of the above corresponding to the field data types data item in the field index of each field.DA can also determine the final position of this field in binary data file according to the data type of each field, thereby determine the reference position of next field in binary data file, and revise respectively in view of the above corresponding to the field address data item in the field index of each field.
After handling first record, thereby DA determines the reference position of the definite next record of length of this record according to the length code of first record of binary data file, begin to handle next record, process and the said process handled are similar, so repeatedly the record of all in handling binary data file.
When the scanning binary data file, DA can calculate the number that writes down in this binary data file simultaneously.
Need to prove that each field index all has identical data length; Equally, each recording indexes also all has identical data length.Like this, in the field index file, just can have access to the field index of specific field correspondence very soon; Equally, in the recording indexes file, also can have access to the recording indexes of designated recorder correspondence very soon.
Based on by the described secondary index structure of Fig. 1, utilize the method for concordance list fast access binary data file below in conjunction with the more detailed explanation of accompanying drawing.Please refer to Fig. 2, Fig. 2 is the process flow diagram of visit data in the access method of binary data file provided by the invention.
Here, described obtaining is meant: when DA carried out data analysis, DA can receive external command and obtain the data type of field to be visited and the data type of this field place record from external command.In the binary data file of a reality, the record with same data type has several.In the common application, visit the corresponding field in all records with specified data type if desired, the data type that provides field data types and field place record by external command gets final product.
Step 204 is determined this address of the field index that writes down first field in the field index file according to the lead-in segment index address data items of current record index.
What deposited the lead-in segment index address of recording indexes is not the address of lead-in section in binary data file of this record, but the address of field index in the field index file of the lead-in section that should write down.
Step 206 judges whether the data type of current field index institute corresponding field meets the field data types of appointment, if execution in step 207 then, otherwise execution in step 209.
Here, DA judges whether the data type of current field index institute corresponding field meets the record data type of appointment, be according to address stored value in the field data types data item of current field index, in the dtd file, find the data type of current field, and compare with the field data types of appointment.
Step 207 is determined the address of field to be visited in binary data file according to the field address data item of this field index.
Step 208 according to the address that obtains in the step 207, reads the binary format data of this field and does subsequent treatment, execution in step 210 from binary data file.
Owing to when specifying the field that to visit, be actually after the designated recorder data type, from all fields of record with this data type, select one and visited by external command.Therefore the data type of certain record and the record data type matching of appointment can not appear, and in should writing down the data type of all fields all with the unmatched situation of the field data types of appointment.That is to say, if in step 203 through judging execution in step 204, with regard to one surely in step 206 through judging execution in step 207, further execution in step 208 and 210.
Before reading the binary format data of this field, also need to determine the length of this field.In order to reach this purpose, at first determine the data type of this field by the field data types data item in the field index of this field; Judge that then it still is the dynamic data type that this field belongs to static data types, if static data types is then directly determined field length according to the definition of data type, if the dynamic data type is then determined field length according to the data type of field self with by the value of other fields of the data type appointment of field self.For example, certain has the field of dynamic data type, and the data type of himself is the variable character type array of element number, and the number of array element is to be determined by the value of a no symbol short field of this field front.Like this, the length of described character type array field just equals the product of the value of the length of character type data and described no symbol short field.
Step 209 reads this field index that writes down next field as current field index, returns execution in step 206.
According to the description to the secondary index table that Fig. 1 provided, the data length of each field index in the field index file is the same.Like this, the address of previous field index in the field index file according to current record adds the data length of field index in the field index file, just can obtain the address of field index in the field index file of a back field of current record.
Step 210 judges whether to handle all recording indexes in the recording indexes file, if execution in step 212 then, otherwise execution in step 211.
For the first time scanning binary data file when setting up concordance list, DA can calculate the number that writes down in this binary data file simultaneously, and obtains the sum that writes down in this binary data file.When visiting this binary data file by concordance list afterwards, DA can safeguard a counter, initial value with counter when handling first record of recording indexes file is made as 0, whenever handles a record value of counter is added 1, and and write down sum and compare.If after handling a recording indexes, the value of counter equals the record sum, and all recording indexes of handling in the recording indexes file are described.
Step 211, the next record index returns execution in step 203 as the current record index in the reading and recording index file.
According to the description to the secondary index table that Fig. 1 provided, the data length of each recording indexes in the recording indexes file is the same.Like this,, add the data length of recording indexes in the recording indexes file, just can obtain the address of recording indexes in the recording indexes file of a back record according to the address of previous recording indexes in the recording indexes file.
Step 212 finishes this visit to binary data file.
Need to prove, in step 202, might not need that first recording indexes begins to search from the recording indexes file, as long as can guarantee at last all recording indexes in can the traversal record index file.When setting up recording indexes for binary data file, the pairing index of record with same data type can also be left in the same recording indexes file continuously, and have the pairing index of record of different types of data with different recording indexes document storages.Like this, when visiting binary data file afterwards, data type according to the field to be visited place record that obtains in the step 201 just can directly obtain the pairing recording indexes of all records that meet this data type, compares one by one and not be used in the recording indexes file.
The above is preferred embodiment of the present invention only, is not to be used to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1, a kind of access method of binary data is characterized in that, sets up recording indexes file and field index file for each binary data file, and this method also comprises:
Steps A is obtained the data type of field to be visited and the data type of place record;
Step B, the pairing recording indexes of record that from the recording indexes file, finds the data type with field to be visited place record to be complementary;
Step C obtains the index of field to be visited in the field index file according to the data type of recording indexes that is found among the step B and field to be visited;
Step D, the index of the field to be visited that obtains according to step C reads field to be visited from binary data file.
2, the access method of binary data according to claim 1 is characterized in that, the pairing recording indexes of record that step B is described to find the data type with field to be visited place record to be complementary from the recording indexes file is:
Step B1, the reading and recording index is as the current record index from the recording indexes file;
Step B2, determine the data type of the pairing record of current record index, and whether the data type of judging the pairing record of current record index meets the data type of field to be visited place record, if meet then think that the current record index is the pairing recording indexes of record that the data type with field to be visited place record is complementary, otherwise think that the current record index is not the pairing recording indexes of record that the data type with field to be visited place record is complementary;
Step B3 reads next bar recording indexes as the current record index from the recording indexes file, return execution in step B2.
3, the access method of binary data according to claim 1 is characterized in that, describedly sets up the recording indexes file for binary data file and is:
The pairing recording indexes of record that will have different types of data leaves in the different recording indexes files, and with the sign of the pairing record data type of the stored recording indexes of each recording indexes file as this document.
4, the access method of binary data according to claim 3 is characterized in that, the pairing recording indexes of record that step B is described to find the data type with field to be visited place record to be complementary from the recording indexes file is:
Will with the recording indexes in the corresponding recording indexes file of data type of field to be visited place record, the pairing recording indexes of record that is complementary as data type with field to be visited place record.
5, the access method of binary data according to claim 1 is characterized in that, step C comprises:
Step C1 is the address with the lead-in segment index address data items of current record index, and the data that read length and be a field index data length from the field index file are as current field index;
Step C2, determine the data type of the pairing field of current field index, and whether the data type of judging the pairing field of current field index meet the data type of field to be visited, if meet then think that current field index is the index of field to be visited, otherwise execution in step C3;
Step C3, the next field index that reads current record from the field index file is returned execution in step C2 as current field index.
6, the access method of binary data according to claim 1 is characterized in that, step D comprises:
Step D1 determines the data type of field to be visited according to the value of field data types data item in the field index of field to be visited, and determines the data length of field to be visited in binary data file;
Step D2 is the address with the value of field address data item in the field index of field to be visited, reads in the binary data file data that begin from appropriate address as field to be visited, and the data length that reads is the data length that obtains among the step D1.
7, the access method of binary data according to claim 2, it is characterized in that, the address value of the record data categorical data item put data type in recording indexes, the data type of the pairing record of the described definite current record index of step B2 is:
From with the appropriate address reading of data of the corresponding dtd file of binary data file data type as record.
8, the access method of binary data according to claim 5, it is characterized in that, field data types data item in field index is deposited the address value of field data types, and step C2 is described to determine that the data type of the pairing field of current field index is:
From with the appropriate address reading of data of the corresponding dtd file of binary data file data type as field.
9, the access method of binary data according to claim 6 is characterized in that, step D1 is described to determine that the data length of field to be visited in binary data file is:
Step D11 determines that according to field data types this field is static data types or dynamic data type, if static data types execution in step D12 then, otherwise execution in step D13;
Step D12 determines the data length of field to be visited in binary data file, execution in step D2 according to the definition of static data types;
Step D13 reads definite needed other information of the length of field to be visited in binary data file according to the dynamic data type definition, and then determines the length of field to be visited in binary data file.
10, the access method of binary data according to claim 2 is characterized in that, further comprises between step B2 and the step B3:
Judge whether to handle all recording indexes in the recording indexes file, if handle then stop this browsing process, otherwise execution in step B3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101243552A CN100357952C (en) | 2005-11-29 | 2005-11-29 | Binary data access method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101243552A CN100357952C (en) | 2005-11-29 | 2005-11-29 | Binary data access method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1851701A CN1851701A (en) | 2006-10-25 |
| CN100357952C true CN100357952C (en) | 2007-12-26 |
Family
ID=37133180
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101243552A Expired - Fee Related CN100357952C (en) | 2005-11-29 | 2005-11-29 | Binary data access method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100357952C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101321177B (en) * | 2008-07-18 | 2011-06-22 | 中兴通讯股份有限公司 | Flat code stream node and flat code stream encoding/decoding method of message |
| CN102609452B (en) * | 2012-01-11 | 2014-12-10 | 上海西本网络科技有限公司 | Data storage method and data storage device |
| CN106933974B (en) * | 2017-02-14 | 2020-02-14 | 浙江大华技术股份有限公司 | Video file index information construction and query method and device |
| US10713232B2 (en) * | 2018-04-16 | 2020-07-14 | Computational Systems, Inc. | Efficient data processing |
| CN111414335B (en) * | 2020-03-20 | 2023-07-14 | 深圳市凯沃尔电子有限公司 | Data file searching method and device |
| CN112148673A (en) * | 2020-09-22 | 2020-12-29 | 北京中房智宝科技有限公司 | Method for solving read-write error of multi-binary system combined binary system file |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6006232A (en) * | 1997-10-21 | 1999-12-21 | At&T Corp. | System and method for multirecord compression in a relational database |
| CN1464436A (en) * | 2002-06-26 | 2003-12-31 | 联想(北京)有限公司 | Data storing and query combination method in a flush type system |
| CN1645365A (en) * | 2004-12-20 | 2005-07-27 | 浪潮电子信息产业股份有限公司 | Method for decreasing DBMS field storing redundance |
-
2005
- 2005-11-29 CN CNB2005101243552A patent/CN100357952C/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6006232A (en) * | 1997-10-21 | 1999-12-21 | At&T Corp. | System and method for multirecord compression in a relational database |
| CN1464436A (en) * | 2002-06-26 | 2003-12-31 | 联想(北京)有限公司 | Data storing and query combination method in a flush type system |
| CN1645365A (en) * | 2004-12-20 | 2005-07-27 | 浪潮电子信息产业股份有限公司 | Method for decreasing DBMS field storing redundance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1851701A (en) | 2006-10-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107807982B (en) | Consistency checking method and device for heterogeneous database | |
| CN101739453B (en) | Method and device for carrying out condition query on database table | |
| CN100357952C (en) | Binary data access method | |
| CN103345469B (en) | The storage of set of numbers, querying method and device thereof | |
| CN101499065B (en) | Table item compression method and device based on FA, table item matching method and device | |
| CN101770367B (en) | Compressing method and compressing device of .NET file | |
| CN105144157A (en) | System and method for compressing data in database | |
| CN112328589B (en) | Electronic form data granulation and index standardization processing method | |
| CN110795464B (en) | Method, device, terminal and storage medium for checking field of object marker data | |
| CN115905630A (en) | Graph database query method, device, equipment and storage medium | |
| CN101963993B (en) | Method for fast searching database sheet table record | |
| CN101388018A (en) | Computer aided design document management method | |
| US8312052B2 (en) | Process for transforming and consulting directed and attributed multigraphs based on the use of maps and bitmaps | |
| CN111859863A (en) | Document structure conversion method and device, storage medium and electronic equipment | |
| CN104750743A (en) | System and method for ticking and rechecking transaction files | |
| CN114185934B (en) | Indexing and query method and system based on Tiandun database column storage | |
| CN108241622B (en) | Query script generation method and device | |
| CN114564501A (en) | Database data storage and query methods, devices, equipment and medium | |
| CN102043856B (en) | Data acquisition method and system | |
| Guo et al. | A study of UHF-RFID data model construction in university libraries | |
| CN113157692B (en) | Relational memory database system | |
| CN114356912A (en) | Method for writing data into database and computer equipment | |
| CN112419058A (en) | Data management method and device, storage medium and electronic equipment | |
| CN100361088C (en) | Method for realizing device test | |
| CN112527745B (en) | Embedded file system multi-partition analysis method, terminal device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20071226 Termination date: 20161129 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |