Execute interface and go to Operator tab.
Expand task node and view execution steps.
Find header-generated step and view code.
If code is empty then step not executed => check IKM code - why.
If code is not empty => check that code correct (not generated empty header).
I had similar issue. Setting all field data types to string in the Datastore did the trick to generate the headers, but introduces other issue with double quotes around numbers. Feature request was placed with Oracle
Ok, getting the following error in step 6 - Integration - HFM_EA_Translate - Insert Column Headers
The last two columns in the file are numeric which seems to be causing the issue. Will change formatting for the two columns (String, Numeric, etc.) to see if I can resolve the header issue.
Ok, converting the numeric columns to string resolved the issue with the header not being generated.
However I now see the real problem that the numeric column has leading spaces in the number to make the the column 30 in length. Seems like a bug with HFM EA extracts and data values.
Now I need to remove the spaces from the data value. May need to create a table to load the data to to use the string functions.
Any other suggestions would be helpful.
Have you tried changing the order of the columns? Moving the numeric columns so they are not last, and keeping their datatype as numeric, and instead placing a string column at the end? I remember seeing this as the solution to a similar issue at some point, but cannot recall the exact details. Worth a shot.
if it is only about removing leading or trailing spaces you could try using a Linux/Unix tool like sed or awk on your source system (if your ODI Agent is running there).
To be honest, I actually think it is more a bug than a missing feature when an ETL tool is not capable of creating proper csv files as i.e. described here: http://en.wikipedia.org/wiki/Comma-separated_values
You can implement the generation in Jython, its quite easy.
vSrc = open('<%=odiRef.getSchemaName( "<YourSchema>" , "D" )%>/<%=snpRef.getTargetTable("RES_NAME")%>', 'w')
vCol = "<%=snpRef.getColList("", "[COL_NAME]", ";", "", "INS") %>" + "\n"
Just add this command to KM.