Length of a character string?

Claes Norreen · Posted: Thu Jun 07, 2012 2:56 pm

Hi,

Is it be possible to find the length of a character string with DFSORT?

I need to load data with the DB2 LOAD utility, and I need this for the purpose of finding the length of a varchar column.

I know it could be done by using the RDW, but I'm looking for a simpler solution. Is there one - or is this feature missing from DFSORT..?

Thanks,
Claes

Claes Norreen · Posted: Thu Jun 07, 2012 4:30 pm

I need to find the length of a character string. I.e.:

XXXXXXXXXX (10)
AAAA (4)
ABCDEFG (7)

By isolating the string in a record that would hold the longest string, I could extract the length from the RDW. However, that's not a very straight forward solution. So I was looking for an easier way to deal with it using DFSORT. But not sure if DFSORT can even do that? If not, I would like it for a future enhancement.

Anuj Dhawan · Posted: Thu Jun 07, 2012 5:43 pm

Claes,

Bill Woodger · Posted: Thu Jun 07, 2012 6:32 pm

Is it just one character string? Is you file FB or VB?

Alternatively, just a big bunch of IFTHENs looking for the trailing pad character of your field, size increasing one at a time (start decreasing) and you have the size of your field.

dick scherrer · Posted: Thu Jun 07, 2012 7:58 pm

Hello Claes Norreen,

Welcome back!

FWIW - Claes Norreen is one of the more senior people who visit here, though it has been a long while since his last visit.

If there is only one variable length field in the file (and the file is vb) the string length might be determined using the rdw.

Skolusu · Posted: Thu Jun 07, 2012 9:44 pm

Claes Norreen · Posted: Mon Jun 11, 2012 1:48 pm

Hi all,

I've been away a few days, but I'm back now. Thanks for your replies.

The field in question has a max. of 250 chars. So, although a bunch of IFTHENs could do it (thanks Bill), thats not really an option.

The fields is just one of many fields. As I say, I know I could isolate the field in a VB dataset, and extract the length of it from the RDW - but I was looking for something more simple.

Kolusu - maybe you can add this to your long list of future enhancements? ;-)

Claes Norreen · Posted: Mon Jun 11, 2012 1:49 pm

Bill Woodger · Posted: Mon Jun 11, 2012 2:07 pm

If it is just the one field you need the length of, and assuming you have fixed-length records, if you just make the field the last one and FTOV with TRIM then the difference in the record-length compared to the maximum length will tell you.

If this does not suit, don't discount the IFTHENs. No-one wants to type out (or even an "EDIT JOB") 250 IFTHENs, changing each one, but you could generate (from DFSORT or some other software of your choice) the IFTHENs with all the changes in. I think this is the route I'd go.

Test with a few bytes so you have all the code you need. Then look for some examples, if choosing DFSORT, of REPEAT. In a new step, USE REPEAT to generate 250 "model" IFTHENs, use INREC or OUTREC to change each on, use a count of 250 to get the closing bracket on the last.

Claes Norreen · Posted: Mon Jun 11, 2012 2:31 pm

Bill, yes that's possible

I could just generate the TOOLIN/xxxxCNTL cards using the software we got here, and get it done that way. Thanks for your input and time.

Though, I still think it would be a nice addition to our beloved tool. ;-)

Bill Woodger · Posted: Mon Jun 11, 2012 2:42 pm

No problem. Good luck. 250X is 250 spaces, so you don't need huge strings across multiple lines wondering about the continuations...

Edit: Unlike my implication, you don't need to generate for each run, of course :-)

If your file is already variable and field has no embedded blanks (unlikely) you could search the DFSORT forum for HUMPHREY.