Adding records (File architecture): Difference between revisions
mNo edit summary |
mNo edit summary |
||
Line 25: | Line 25: | ||
##then, for each field value pair in the STORE RECORD statement (none are required, an 'empty' record can be created if no fields are referenced): | ##then, for each field value pair in the STORE RECORD statement (none are required, an 'empty' record can be created if no fields are referenced): | ||
###the field value pair is added to the record (or preallocated field updated) | ###the field value pair is added to the record (or preallocated field updated) | ||
###of course, normal extension record rules apply (as discussed in [[Table B (File Architecture)|Table B]]) | |||
###if the field is indexed, the necessary updates occur to Tables C and D. | ###if the field is indexed, the necessary updates occur to Tables C and D. | ||
Revision as of 21:02, 3 April 2013
Adding a record to a file
This section describes the process to add a record to a Model 204 file. It presumes that the fikle has already been CREATEd and INITIALIZEd, and that all fields and Repeating Field Groups (RFGs) needed in this action are defined to the file.
And a reminder: while fields and RFGs may, generally, be added to the file at any time, some field attributes (for example, the pre-allocation of fields with the OCCURS attribute), may only be created in a file to which no data has yet been added.
The addition of the record is driven by the User Language STORE RECORD statement with an (optional) list of fieldname value pairs, as in the following example.
STREC: STORE RECORD NAME=RASPUTIN SEX=MALE OCCUPATION=MONK END STORE
Alternatively, you can use the Host Language Interface IFBREC, IFPUT, and IFSTORE functions, which also require a list of name = value pairs.
All referenced field names must previously have been defined, but not all field names need to have assigned values in every record. In fact, if you omit a field, the field is not stored in the record at all, thus saving space in the file.
For more information see the Rocket Model 204 User Language Manual or the Rocket Model 204 Host Language Interface Reference Manual.
How Model 204 adds records to a file
In the process of adding a new record to a file, Model 204:
- Like all User Language code, the process consists of two parts; first the code is compiled:
- all field names, the values of coded fields and some FRV fields are converted into fixed-length codes from Table A
- all other attributes of the referenced fields are noted for use during the execution
- Presuming the code compiles, in the evaluation process, when the STORE RECORD statement is executed:
- a physical internal record number (IRN) is assigned in Table B taking the BRESERVE parameter into account
- if there are any preallocated fields, the fields in the Table D record map are added to the record
- then, for each field value pair in the STORE RECORD statement (none are required, an 'empty' record can be created if no fields are referenced):
- the field value pair is added to the record (or preallocated field updated)
- of course, normal extension record rules apply (as discussed in Table B)
- if the field is indexed, the necessary updates occur to Tables C and D.
Considerations when adding many records to a file
If you are adding a large number of records to a file, consider these alternatives for increased efficiency:
- Deferred updates. Defer changes or additions that update Tables C and D, instead specifying that step 3 above be performed in separate phases. Deferred Update Feature describes how to use deferred updates.
- File Load utility. If the raw data for the records to be stored in the Model 204 file already exists in a machine-readable format (magnetic tape, direct-access storage, cards), use the File Load utility to load that data directly into the Model 204 file quickly and economically. File Load Utility describes the File Load utility.