How to solve Character set conversion error!

User_YGZU0

I am importing a data dump of 11g database in 19c version using Impdp command. During the import I am getting the following message:

"Master table "SYS"."SYS_IMPORT_FULL_17" successfully loaded/unloaded

import done in AL32UTF8 character set and AL16UTF16 NCHAR character set

export done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set

Warning: possible data loss in character set conversions"

Due to this error I am not able to import some of very important tables and it is showing the following error:

"ORA-02374: conversion error loading table "DWH"."GBLT_LOGIN_UNSUCESSFUL_LOG"

ORA-12899: value too large for column GSTR_LOGINID (actual: 28, maximum: 20)

ORA-02372: data for row: GSTR_LOGINID : 0X'554443483137303120BFBFBFBFBFBFBFBF204361'"

I believe the above error is due to character set conversion error. Kindly suggest any solution!

evgenyg

if you can (and want to) touch the source then, in my opinion, the DMU can help there (you will have export and import using the same character set.). If you can't touch the source system then you may want to precreate destination tables and try to import again. In fact the solution is depends on the requirements. I am just pointing out why you, most probably, have the issues initially.

Regards

evgenyg

I think that

WE8MSWIN1252  is not strict subset of AL32UTF8  and before dealing with this kind of issues you may want to read:

https://docs.oracle.com/en/database/oracle/oracle-database/12.2/nlspg/character-set-migration.html#GUID-8F973339-3D66-4C6D-A045-249A81007092

I guess GSTR_LOGINID  is char(20 bytes) in above documentation there is example why you can fail:

For example, ä (a with an umlaut) is a single-byte character (E4) in WE8MSWIN1252, but it becomes a two-byte character (C3 A4) in AL32UTF8. 

You may want to run character set conversion on the source or pre-create that table on destination with properly defined datatypes.

User_YGZU0

Can this problem be solved by using  Oracle Database Migration Assistant for Unicode (DMU)? Or by changing character set using any other method?

evgenyg

if you can (and want to) touch the source then, in my opinion, the DMU can help there (you will have export and import using the same character set.). If you can't touch the source system then you may want to precreate destination tables and try to import again. In fact the solution is depends on the requirements. I am just pointing out why you, most probably, have the issues initially.

Regards