After DB migration(Informix to Oracle), SQLs need to review & change?

Question

Dear all,
We migrate our system DB from Informix to Oracle 9i.
The current problem is: All data in CHAR fields will have "SPACE".
For example,
There is a table named TABLE1, and one field named FIELD_1 in it.
The type of FIELD_1 is CHAR, 4 characters.
In Informix, the data in TABLE1 is
FIELD_1
===============
CM
ALL
TEST
After migrate to Oracle, the data "looks like the same".
FIELD_1
===============
CM
ALL
TEST
But, when using TOAD or SQL/PLUS to query the length of these data, like:
SELECT LENGTH(FIELD_1) AS LEN, FIELD_1 FROM TABLE1;
And, the return result is:
LEN FIELD_1
==== ==============
4 CM
4 ALL
4 TEST
All length is 4..........Orz......
So, all SQL in current AP need to review and change(if necessary).
for example, the statement likes:
WHERE FIELD_1 = 'CM' will change to WHERE TRIM(FIELD_1) = 'CM'
Does anybody have the same probelm?
And another question is, if we always use TRIM in the WHERE statement, will cause bad performance or not?
(because I try to change one SQL and get the return result.....takes more time than before)
Thanks a lot.
Best Regards,
Claire

Answer 1

The current solution is review all codes in AP and change the WHERE statement.Don't do that! You'll have all kinds of problems and any new code you develop, you'll have to remember this hack. Instead, fix the tables once so they are correct. Continuing my example:
SQL> alter table t modify (object_name varchar2(30));
Table altered.
SQL> update t
  2  set object_name = trim(object_name);
7852 rows updated.
SQL> commit;
Commit complete.
SQL> select length(object_name), count(*)
  2  from t
  3  group by length(object_name);
LENGTH(OBJECT_NAME)             COUNT(*)
                   1                    1
                   3                   22
                   4                   14
                   5                   33
                   6                   46
                   7                  213
                   8                  283
                   9                  266
                  10                  295
                  11                  288
                  12                  291
                  13                  339
                  14                  354
                  15                  367
                  16                  428
                  17                  396
                  18                  368
                  19                  377
                  20                  348
                  21                  340
                  22                  353
                  23                  306
                  24                  347
                  25                  320
                  26                  325
                  27                  269
                  28                  332
                  29                  299
                  30                  232
29 rows selected.For each table, you can do one modify and one update statement to fix it. With a little clever use of the data dictionary, you might even be able to script it.
Much easier than hacking your code.
Message was edited by:
Eric H