etl - Buffer length vs column precision vs DTM buffer size in Informatica

Question

I have a workflow which loads millions of data from flat file to Oracle Table. The buffer length set was 10240. Recently it failed with the error record length is larger than the line sequential buffer length error. People suggesting to change the larger columns precision value instead of changing the buffer length in the workflow. My question's are -

1. Why to change only precision value instead of buffer length?
2. Do Increasing the precision or buffer length will affect other sessions running in parallel?
3. Is there any suggestions what to do for such buffer length issues?
4. What exactly DTM buffer size doing here? Doesn't it need to change?

question from:https://stackoverflow.com/questions/65839133/buffer-length-vs-column-precision-vs-dtm-buffer-size-in-informatica

Answer 1

Line sequential buffer length in informatica is a session property which specifies the length of bytes of a row from flat file source. You need to increase this to avoid such error. it seems you are getting data longer than usual. Answer to your questions-

1. Changing precision wont help because infa still need to read whole row from FF.
2. Increasing buffer length will need more memory so may be it will impact other parallel sessions.
3. Apart from increasing Line sequential buffer, you can also add memory to the default buffer block size and DTM to enhance perf (it will impact other sessions running in parallel ).
4. DTM buffer applies while data movement, your issue seems occuring while reading the data form file so DTM isnt going to help.