There are no sort keys in the dataset schema

[ravij]

Hi All,

My job design is: Orastage-->Sort-->Trf-->dataset

I haev two sort keys defined in Sort stage. when I run the job, am getting warning message like:

main_program: Sort key "IIMS_REF_NO" no longer exists in dataset schema. It will be dropped from the inserted sortmerge collector
main_program: Sort key "CE_BATCH_ID" no longer exists in dataset schema. It will be dropped from the inserted sortmerge collector
main_program: There are no sort keys in the dataset schema. No sortmerge collector will be inserted.

I searched forum and found related threads but couldn't find required solution. I am just using Auto partition in all stages.
Please help me in removing the warnings.
Thanks in advance.

[ETLJOB]

I am just using Auto partition in all stages.

Try to specify "ordered" or "round robin" collection methods in dataset and check the warnings are existing or not. I hope "sort merge" collection method automatically imposed by DataStage behind the scenes as you have a sort stage in the flow.

[ravij]

Hi folk,

Thanks for your response.
I could not find "Ordered" in the list and tried with "Round Robin" in dataset stage(target in my job) partition tab, still getting the same warning along with one more new warning like

Data_Set_122: When checking operator: Operator of type "APT_CopyOperator": will partition despite the
preserve-partitioning flag on the data set on input port 0.

Any help would be appreciated

[igorbmartins]

Ravij do you need to do this:
Transformer Stage Properties; Stage; Advanced; Preserve partitioning;Clear

[ravij]

Hi igorbmartins,

I tried with that option, but no luck. It was not working and the same 3 warning messages.

any other help please.

Thanks in advance.

[madhukar]

check for below
-renamed sort columns in output link of sort stagekeeping column names
-check if you have set clear partition anywhere before sort stage

[ravij]

Hi Madhukar,

check for below
-renamed sort columns in output link of sort stagekeeping column names
-check if you have set clear partition anywhere before sort stage

I checked both the cases which you mentioned, it was default(clear) in orastage(which is before sort stage in my job).
And the other one is, the two column names same in Input link and Output link in Sort stage.
I am getting these warnings only for few jobs rest are all running fine. Even i cross checked with the other jobs which are running fine..there is no change in design and properties etc..

Your help would be greatly appreciated.

(fixed oddly quoted post for you - Moderator)

[sauros73]

Hello everyone,

I had that issue and try to play with partitionning (Hash, Round Robin, etc), add a Sort stage, etc. nothing worked.

NOW I FOUND iT !!!

here my job stages design:

2 Datasets (parallel) > Change Capture > Transformer > 5 Sequential files

My 2 dataset have the same data definition and the same key but my 5 Sequential files didn't not have that key for all of them.

The problem was not at the Change Capture stage but into the transformer because the option "Preserve sort order" was checked for my outputs (to the sequential files) and some of these files didn't not have the sort key.

Here how you find that option

Open transformer > open its properties > tab Outputs > for each output links, uncheck the option (left bottom of the window)

Hope it will help you

[ray.wurlod]

Merci, Hugo.