NLS mapping error

kab123 · Post by **kab123** » Tue Mar 22, 2005 2:49 pm

Hi All,

Can I get any help how to suppress this warning error.
"nls_map_buffer_out() - NLS mapping error, row 12345 row =......"
I dont see this error for all jobs.. The jobs runs ok, but I dont want to see these warnings too..

Thanks

ray.wurlod · Post by **ray.wurlod** » Tue Mar 22, 2005 11:17 pm

Please post the full text of one of these messages; otherwise it is impossible to diagnose whether it's important or not. It probably is; you've possibly lost some data because one or more characters could not be accurately mapped. When does this error occur? In the log, or only when you are using the data browser?

kab123 · Post by **kab123** » Thu Mar 24, 2005 10:46 am

Ray,

JobName:.link.stagename:"nls_map_buffer_out() - NLS mapping error, row 12345 row = 'the whole record'..

The above given is the full message I could see..and this is seen in log and not when I view the data.. and I could see the same number of records loaded.. so i dont have any missing data...this error gives when i write to a file and when I read from same file it doesn't give any of these warnings..
I hope this info helps...Is this something need to do in NLS Tab in Seq file....:?:

Thanks Ray...

ArndW · Post by **ArndW** » Thu Mar 24, 2005 10:52 am

Kab,

what is most likely happening is that you are getting an input character that has no corresponding output character in the destination NLS map. Can you tell us what your input and output sets & maps are? If you are dealing mainly with single-byte ascii type data, perhaps you could look at the row in question in an editor and see if any "funny" bytes show up - this would help localize the error cause. As Ray stated, you will most likely be losing these characters.

ray.wurlod · Post by **ray.wurlod** » Thu Mar 24, 2005 4:17 pm

Are there any "?" characters in "the whole record" in the logged message? These are used to indicate which characters can not be mapped.

kab123 · Post by **kab123** » Fri Mar 25, 2005 1:13 pm

This is the actual log error...
I dont see any "?" in the record. but I am getting Null which is shown as "" in the log..It is unable to handle the nulls..do I need to set something at Default NULL String in the target seq file..?
nls_map_buffer_out() - NLS mapping error, row 16858 (approx), row = ""123456789","279522881","33645","05/17/1994","","ADAMS","PATRICIA"," "," ","MD","800 MIDLAND AVE"," ","MEMPHIS","TN","38123","9015457487","DEPT 150 PO BOX 300"," ","MEMPHIS","TN","381010357",,"HEMATOLOGY / ONCOLOGY","","","","",15"...

Thanks

ray.wurlod · Post by **ray.wurlod** » Fri Mar 25, 2005 4:38 pm

It's certainly worth a try.

You can set it for just that column; scroll to the right in the Columns grid of the Sequential File stage's Inputs tab Columns page to see how.

kab123 · Post by **kab123** » Wed Apr 20, 2005 11:53 am

Ray,

Yes as u said, I could see "?" in the record...where I get nulls.
In my source file which is a comman delimited..doesn't have quotes..so, for nulls It is wrriten as "?"...And I get lot of such warnings...
but in the log, the warning is shown as :
nls_map_buffer_out() - NLS mapping error, row 75 (approx), row = "58-1972773,,,,3/1/2000,,A,G,,Healthcare Corporation of Northern California,,,,,,3160 Corporate Place,,Hayward,CA,94545,5107328800,,,Healthcare Corporation of Northern California,PO Box 74798,,Chicago,IL,60694-4798,9166148979,,,Home Infusion,,,,,,,,,, "

Any suggestions..

thanks

kab123 · Post by **kab123** » Wed Apr 20, 2005 1:54 pm

Nulls are not read properly and write as nulls...
how to handle nulls properly..?

kab123 · Post by **kab123** » Wed Apr 20, 2005 2:26 pm

only some Nulls are not readable and inturn they are written as "?" in the tgt fil.....but not always all nulls are written as "?".
How to handle this NLS mapping warnings..? :(

Any help pls...

Thanks in Advance

ArndW · Post by **ArndW** » Thu Apr 21, 2005 12:09 am

kab123,

if, in some rows, NULLs are handled correctly and in others they aren't then it would be logical to assume that they are not the same character, even though they appear to be. NULL is mapped in just every character set and mapping that I know of.

Can you put a second output link to your transformer and output that one row you are getting from a derivation of

Code: Select all

SEQ(In.ColumnName)

? Then look that character up in your ASCII table or tell us what the value is? Plus tell us what your NLS character set is for the job or column in question?