Re: parsing a CSV column

On 07-Aug-2015 11:00 -0600, rob wrote:

Lets say I run a command the automatically populates a db2 column
with:

lparname1,1,status
lparname2,2,status
lparname3,3,status

No danger of a comma being imbedded other than as shown. IOW no risk
of: "lpar,1,name",1,status.
I know I can parse this out with SQL with a bunch of LOCATE's and
whatnot. But, if there's some cool function that I'm missing I'd
rather examine that.
My workaround may be to pipe this out to a stream file instead of a
DB2 file and use CPYFRMIMPF. And if "I'm" looking at CPYFRMIMPF you
know that's desperation.

No CSVPARSE(fromstring, ColumnNbr) is there?

I know of no built-in scalar or table function to effect that in DB2; the following discussion suggests the INSTR function in Oracle/PLSQL is capable: [http://stackoverflow.com/questions/29983259/select-a-portion-of-a-comma-delimited-string-in-db2-db2400]

With my response in that discussion, I had offered a Recursive Common Table Expression (RCTE) for effectively the same intent [break the comma-separated tokens] in another discussion here, found at [http://archive.midrange.com/midrange-l/200706/msg01329.html], and that could be used as the basis for implementing either variation of the User Defined Table Function (UDTF) to which I had alluded as a parsing function.