The silence here probably means the audience is not familiar with "COBOL program where the field is 9(7).9(7)". Anyhow, I don't underdstand. I can say however that in your specs you have two things that are not required, hence pure overhead: "PAD BLANK" is default and when padding is with blanks, there is no need to code things like "/ / 105" (unless the data you placed previously could have placed something else in column 105 & 106).
2007/11/15, [EMAIL PROTECTED] <[EMAIL PROTECTED]>: > > > i am using pipelines to reformat a DB data dump file that has vertical > bars as field separators. > > one of the fields is a dollar amount which can take any of these formats: > 0 > 1 > 12 > 123.4 > 1234.56 > > i want to reformat this as input for a COBOL program where the field is > 9(7).9(7) > > can i do this in the specs stage ? > or do i have to pass the records thru and format this particular field > and then use my specs stage to put all the fields into the right columns? > > here is the current coding i have: > 'PIPE', > '< CABSEAST DATA B', /* read the oracle dump file */ > '| drop 1', /* skip the first line - header data */ > '| SPECS pad blank fs 4F ', /* fields separated by '|' character */ > 'field 1 1 left', /* field 1 will go into col 1 */ > 'field 2 9 left', /* field 2 will go into col 9 */ > > ... etc. > > 'field 17 104 left', /* field 17 will go into col 103 */ > '/ / 105', /* put two blanks in 105-106 */ > 'pad 0 field 18 107-120 right', /* field 18 is numeric - right align */ > '/ / 121', /* insert a blank */ > 'pad blank field 19 122 left', /* field 19 will go into col 122 */ > 'field 20 124 left', /* field 20 will go into col 124 */ > > ... etc. > > field 18 is the one that needs to be reformatted. > > > prg > > Phillip Gramly > Systems Programmer > Communications Data Group > Champaign, IL -- Kris Buelens, IBM Belgium, VM customer support