I've got case where I dump a table from Oracle 8.1.7 using Oracle 10 client and Bigint columns get corrupted, the corruption causes a key column to apear in several rows in my output data (not entire duplicate rows just the key bigint column gets messed up)
![Shocked :shock:](./images/smilies/icon_eek.gif)
I've used sqlplus with both 8 and 10 (lib32) Oracle clients to dump the data but no duplicates for that key were found
![Sad :(](./images/smilies/icon_sad.gif)
Also if I specify a where clause to get that specific key in my Oracle stage I only shows 1 row.
I've found this due to investigation of deadlocks occuring on job that performs upsert to a table (If ayone is interested).
i.e.
col1 bigint
col2 bigint
col3 char
col1 + col2 are pk
actual values
col1 col2 col3
111 222 abd
112 222 vdsv
Unloading to sequential file using DS parallel job gives
111 222 abd
111 222 vdsv
Anyone had this before
![Question :?:](./images/smilies/icon_question.gif)
Any help will be helpfull
![Exclamation :!:](./images/smilies/icon_exclaim.gif)
Thanks,