NCL File Processing › Work with UDBs › Restrictions When Using UDBs
Restrictions When Using UDBs
When designing facilities that will use file processing the following must be taken into account:
- All records must be keyed, unless processing with an ESDS.
- Keys must range in size from 1 to 255 bytes.
- Keys for base clusters must start in position 0 (unless processing in unmapped mode) of the record. This is often termed relative key position (RKP) 0.
- Allowance should be made for field separators, or length and tag bytes added by NCL when determining record sizes for mapped or delimited files.
- The maximum record size is 32K when doing I/O from NCL variables. Otherwise it is the maximum record size for the file.
- When retrieving records, fields are returned in variables which have a maximum length of 256 characters. Creating fields in excess of this length might not be practicable, although a field in excess of 256 characters can span multiple variables after an &FILE GET.
- Unmapped format UDBs have no record structure maintained by NCL. The user is responsible for determining the record format of unmapped format files.
- Mapped and unmapped format UDBs can contain non-character data. The user must make allowance for this and might want to convert the data to expanded hexadecimal format after it is retrieved from an unmapped format UDB.
- If a UDB is of a format describable to Mapping Services using a map, then mapped format processing can be performed on it.