DSXchange

Hi, This simple programming task made me spent couple of days trying to implement in ds, to no avail :oops: If I have this data: col1 col2 id1 5 id2 0 id3 0 id4 0 id5 0 I'd lke to output to become just one row, where col2 is the last non zero value read. col1 col2 row 5 My approach is compare curren...

so i used an execute command stage, to test the filesize, if it's zero, do something, else, some else.

thanks for the inputs!

hi, i have a lookup stage that if there is a failure, a log will be inserted in the error log file. however, even if there is no lookup failure (reject), the log file is created although it's empty. is this an expected behaviour? i don't want the file being created if there is no lookup failure. TIA!

hi, there as so many links/samples about DSLogWarn. but i could not find any how/where to enter the code? it seems like it can only be done on a server job? what of with parallel? i have a lookup that is set to "Reject", but i want to write a custom warning message in the log as well. TIA

preparing, like updating old records before new ones come in.

thank you for your inputs.

there is no before/after sql. this job is actually part of a sequence. it's like preparing the table before the main job does it's thing.

ok. i must mention RG is creating one dummy record just to have an input to next stage. because it will not work having just db2 stage there by itself. this design just feels a little awkward.

hi,

i needed a job to update records on the db. so i created

row_generator -> db2_connector

above works, but, how would you design the job differently?


TIA!

thank you so much for the tips.

hi,

the op in this link pretty much describe what i need.

viewtopic.php?p=313299

however, i could not find execute command stage in my pallet.

any idea how to check if file exists, and rename afterwards before running the job runs?

i was just citing an example. but thank you. indeed i will try Paul's suggestion.

You mean Transaction - Record Count? It was at default, 2000. I changed it to value higher than the number of records being processed. Still no good. (Although, I would not do this as the number of records is unknown).

hi,

I have a look that is set to FAIL on lookup failure. However, if there were already written records before the failure, they are not reverted.

For example if I have 10 records, and 6th record fail, the job aborts indeed, but the previously written 5 records stays in the database.


TIA!

Hi, Incoming data have unique key values. Auto Commit : off Array Size: 2000 Transaction Record Count : 2000 to test, changed to: On 1 2000 it worked! thank you. Tried different numbers for Array Size and Record Count, all tests failed except only when Array Size is 1. It's running painfully slow ho...

Hi, I have a stage where write mode is Delete and Insert. If I run the job, I would get a warning that says the Delete "failed to run", followed by fatal error "QLSTATE = 02000: Native Error Code = 100: Msg = [IBM][CLI Driver][DB2/NT64] SQL0100W No row was found for FETCH, UPDATE or D...