DSXchange

Here is our job design

Teradata Enterprise -------> |
------------------------------------| Change Capture ------> Output
Sequential File----------------> |

Here data from Teradata Enterprise and Sequential file are identical, but CDC gives most of the record as new records.

"main_program: When preparing operator: When setting up transfer for "LN1_ADR":Implicit conversion from source type "ustring[max=60]" to dest type "string[max=60]": Converting ustring to string using codepage ISO-8859-1."

We identified above warning and we tried changing extended property to "UNICODE" and it works fine as it suppose to be.

Following is "tera_cs.txt" 18 lines, 662 characters

ASCL_LATIN LATIN1_0A
ASCL_EBCDIC EBCDIC
ASCL_KSC5601 HANGULKSC5601_2R4
Shift_JIS KANJISJIS_0S
ASCL_SHIFT-JIS KANJISJIS_0S
ASCL_JPN-EUC KANJIEUC_0U
ASCL_MS1252 LATIN1252_0A
ibm277 EBCDIC277_0E
ASCL_EBCDIC-IBM933 HANGULEBCDIC933_1II
ASCL_EBCDIC-IBM937 TCHEBCDIC937_3IB
ASCL_EBCDIC-037 EBCDIC037_0E
IBM273 EBCDIC273_0E
ASCL_EBCDIC-1026 KANJIEBCDIC5026_0I
ASCL_ISO8859-9 LATIN9_0A
ASCL_GB2312 SCHGB2312_1T0
ASCL_BIG5 TCHBIG5_1R0
ibm-1047-s390 EBCDIC
ISO-8859-1 ASCII

We have tried APT_TERA_NO_SQL_CONVERSION environment variable option also.

Does anyone have faced this problem?

Is there any better way of solving this problem instead of changing the metadata to UNICODE?

Any response on this will be greatly appriciated.

You could convert the Unicode data to non-Unicode in an upstream Modify stage. Add a NOWARN specification.

But what's wrong with performing a Unicode comparison?

ray.wurlod wrote:You could convert the Unicode data to non-Unicode in an upstream Modify stage. Add a NOWARN specification.

But what's wrong with performing a Unicode comparison?

Thanks for your reply Ray

I thought we could have small little configuration parameter in the datastage, so change would be only in that configuration level rather than doing job by job change.

We are not handling data from multiple NLS as well.

No such configuration parameter of which I'm aware.

ray.wurlod wrote:No such configuration parameter of which I'm aware.

This is so difficult to accept that we have to do conversion from Unicode to Ascii by modify or transformer before we use them in the join.

I guess people worked on teradata should have experienced this issue...