in server Sequence file stage there is a property called Defaut NULL String , what is same property in
Parallel Extender
and also incoming data from sequence file is coming with dot (.) i want this should convert into null
default , in coming column field is numeric.
what could be the possible solution for this question.
Sequence file stage Defaut NULL String property
Moderators: chulett, rschirm, roy
-
ray.wurlod
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
As with the server equivalent, you have a null value replacement property in the stage properties. It is an optionally added property - you will find it in the bottom right pane when record-level properties are selected.
As with server jobs you can override this on a per-column basis. Open the Columns tab, select the row on the grid that you want to change, right mouse click and Edit Row. You should find the property readily enough.
Beware that the null-representing value (or the null-replacement value) must have the same width as the width of a fixed-width column. Parallel jobs are finicky about data types being correct.
As with server jobs you can override this on a per-column basis. Open the Columns tab, select the row on the grid that you want to change, right mouse click and Edit Row. You should find the property readily enough.
Beware that the null-representing value (or the null-replacement value) must have the same width as the width of a fixed-width column. Parallel jobs are finicky about data types being correct.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Veni,
that is the difference - the null in the defaults is not an oracle null value. You will need to put this conversion in a modify stage or a transform. I think it will be best to keep the period to 0x00 conversion and then use the handle_null in a modify stage. An alternative that is not quite as efficient would be to remove your handling of the input and then use the CONVERT function in a transform stage, converting the period to char(128).
that is the difference - the null in the defaults is not an oracle null value. You will need to put this conversion in a modify stage or a transform. I think it will be best to keep the period to 0x00 conversion and then use the handle_null in a modify stage. An alternative that is not quite as efficient would be to remove your handling of the input and then use the CONVERT function in a transform stage, converting the period to char(128).
</a>Thanks for your immideate response
Is there no any other possible way because my source is having 1220 field out of that 700-800 filed are numeric (where i have to convert dot (.) to oracle NULL )so using modify or transformer stage is tedious process.
In Modify stage i can't bring 700-800 fields or i can use trim and SetNull function to replace dot (.) in transformer .Is it make sence in term of performance ?
Is there no any other possible way because my source is having 1220 field out of that 700-800 filed are numeric (where i have to convert dot (.) to oracle NULL )so using modify or transformer stage is tedious process.
In Modify stage i can't bring 700-800 fields or i can use trim and SetNull function to replace dot (.) in transformer .Is it make sence in term of performance ?
rv
If your source is a sequential file, define it as having 1 column (with a very long string) and no delimeters, read it and perform the conversion of all "." to CHAR(128), write it back, then read it as before. Or use any one of many different UNIX level options to effect the same change to the file.
Sure it is tedious to put a handle_null in all the columns, but you could always do just a couple, export the job as a .dsx file to see how it is done and use an editor to manually insert all the other handlers. Or you could use a transform stage which will let you perform like operations across columns.
Sure it is tedious to put a handle_null in all the columns, but you could always do just a couple, export the job as a .dsx file to see how it is done and use an editor to manually insert all the other handlers. Or you could use a transform stage which will let you perform like operations across columns.