DSXchange

DELETE.FILE won't help after you have deleted the dictionary and hash file manually. You have to edit the VOC file manually. Hope someone helps you out on that. Sorry.

Hope the hashfile was externally pathed and in case if it was not you used DELETE.FILE command when you deleted it.

Is your oracle date format is in MM/DD/YYYY?. Verify what format your oracle DB accepts. Usually it is in YYYY-MM-DD format. HTH.

Did you try using LongVarBinary?

Hi dls,
As I have always seen IPC using /tmp can you please tell how you could reason out that its conditional?

Thanks.

There should not be much difference between connecting to Oracle using ODBC or sqlnet in my IMHO. Is there a change in your query between now and then?. Or try changing the Array size to a larger value in OCI stage. HTH.

Thanks for the replies. UniVerse does not have a TimeStamp data type. Hashed files store all data as string Thought so.. well, was just a thought :roll: ... Decided to change the job design, as my UV knowledge is limited!. Changed the Universe stage as the driver and OCI as reference and now I have ...

Hi, Does using Universe stage to query records in a hashfile reduce the processing speed?. I used the universe stage to write a user defined query. The hashfile has two keys empId and Date and the query would return if passed-in date is greater than the hashfile date for empID. Now the rows per seco...

You might find THIS similar post helpful.

Certainly can do. But verify whether the integrity of the underlying table is not broken when you update the key field itself.

There should be a source stage somewhere in your job before the transformers. Anyways why don't you try in some other client machine? (if you couldn't pull up the frame back into view that is)

Neeraj, The query that Ken has posted and results thereof should be correct. What is the second date in your output's second row - 100 2 11-MAY-05 10:22:01.000000 AM 11-MAY-05 11:23:01.000000 AM Your source doesn't have a seq 2 for NBR 100 and EVNT_CD 6000 but your result seems to have it. Or otherw...

Reference lookups using OCI, ODBC or Hashfiles. Do a search for outer join.

Seems like you are still connecting to the database using ODBC, from the log. IMHO whenever you see this "Optional feature not implemented" message from ODBC then you have not specified the correct column metadata. Reload it from the table definitions.

That would sure work!
If you want you can break it down to individual lookups into attendance table and derive those counts for each customer id. But performance would be poor compared to Ken's suggestion.