Retrieving Data With SQL › Retrieving Floating Point Data from SQL Columns

Retrieving Floating Point Data from SQL Columns

Although CA Culprit does not support real and double precision as REC parameter data types, there is a simple coding technique to work around the deficiency.

CA Culprit will treat real SQL columns as four byte alphanumeric fields. Double precision columns will be treated as eight byte alphanumeric fields. 16 byte packed decimal work fields can be defined, and CULLUS36 can be used to convert the floating point values to packed decimal. The CA Culprit program below illustrates the technique.

Converting to Packed Decimal

 col. 2
 ▼
 IN DB(Q)  DICTIONARY=TSTDICT   SCHEMA=INV
 SQL SELECT * FROM FLOAT_TABLE;
$  CA Culprit will generate  the REC cards shown below.
$REC  REAL_DATA     49  4                      $ GENERATED
$REC  DOUBLE_DATA   53  8                      $ GENERATED
$
 010  WK_REAL    0.000000
 010  WK_DOUBLE  0.000000000000
 010  RETCODE '    '
 0151*001   WK_REAL   HF
 0151*002   WK_DOUBLE HF
 017      CALL US36(REAL_DATA,  'S',  WK_REAL,    6, RETCODE)
 017      IF RETCODE NE '    '  200
 017      CALL US36(DOUBLE_DATA, 'D', WK_DOUBLE, 12, RETCODE)
 017      IF RETCODE NE '    '  200
 017      TAKE
 017200   CALL US48(' ERROR IN FLOATING POINT CONVERSION%%')
 017      DROP

The arguments for CULLUS36 are described below:

US36 Arguments(Error: Don't know what to do with column width of "20," (3).

ARG1 field-name

Name of input field containing the floating point value

ARG2 indicator

'S' = Single Precision:

'D' = Double Precision

ARG3 result-field

Name of 16 byte packed decimal work field

ARG4 decimal-pts

Integer indicating number of decimal places in result field

ARG5 return-code

Name of four byte alpha work field