C1, C2, C3

Type C is used to define groups of fields whose byte values have similar frequency distributions. Up to 3 different frequency distributions can be accommodated, one each by type C1, C2 and C3. If no other type code is clearly preferable, choose a type C.

Compression is variable, depending on the skewedness of distribution. The more greatly skewed the distribution, the greater the compression ratio. Processing overhead is minimal (3).a

CS

Type CS is used to define groups of fields where the data varies considerably. Compression is variable, according to the data characteristics. Processing overhead is minimal (3).

GA

Type GA is used to eliminate unneeded fields from the compression record.

Compression is 100 percent. Processing overhead is negligible (1).a

L

Type L is used to insert a binary tally of the compressed actual number of bytes comprising the compressed record as the first 2 bytes of the record following the RDW. Particularly useful for COBOL users.

This field type actually increases the compressed record length by 2 bytes. Processing overhead is negligible (1).a

MA, MB

Type M is used when data in a field repeats from record to record. Several restrictions govern use of this field type. For fields smaller than 10 bytes in length, type C is preferable.

Compression is variable, depending upon the degree of data repetition. Processing overhead is variable, decreasing as field length increases (3–5).a

N

Type N is used to exempt a field from compression. Use for retrieval, match and sort keys, and any field which you want to access without expanding the data. There are several restrictions governing use of this field type specification.

This field type yields no compression. Processing overhead is negligible (1).a

PD

Use type PD for fields containing packed decimal data, preferably with many high-order zero digits. If significant digits frequently fill the field or if invalid data is frequently present in the field, choosing a type C specification is preferable.

Compression is excellent when the value is zero, and variable, increasing as the proportion of significant digits to total digits decreases. Processing overhead is moderate (5–6).a

S, X

Use type s or x when the number of values occurring in a field is small (16 or fewer) and these values are known in advance.

Compression is excellent, and the processing overhead is minimal (3).a

UN

Use Type UN for fields which cannot readily be compressed (for example, bit switches, floating point numbers), particularly when all 3 type C specifications are already in use.

This field type yields no compression. Processing overhead is minimal (2).a

V, VP, VZ

Use one of these field type specifications to define a field whose content is used to calculate the actual length of a variable-length portion of the record, or a field which contains the actual length of a variable-length portion of the record.

Type V is not compressed. Type VP is compressed as PD. Type VZ compressed as ZR. Processing overhead is minimal to moderate, depending on field type:

V=2 VZ=4 VP=6a

ZL, ZR

Type Z is used for fields containing zoned decimal data, particularly when a type C specification cannot be dedicated to zoned decimal data. Several restrictions govern the use of this field type specification.

Compression is excellent when value is zero, variable, increasing as the magnitude of the value increases. Processing overhead is moderate (3–4).a