I am using 11g Release 1 Database .
I have to analyze the performance of two tables of different designs which serve the same purpose and come up with the design which is efficient .
SQL> desc staging_dict
Name Null? Type
----------------------------------------- --------
SNO NUMBER
CODE_FRAGMENTS CLOB
CODE_FRAGMENTS_U CLOB
CODE_FRAGMENTS_D CLOB
CODE_FRAGMENTS_DO CLOB
SQL> desc staging_dict1
Name Null? Type
----------------------------------------- --------
SNO NUMBER
CODE_FRAGMENTS CLOB
CODE_FRAGMENTS_UD CLOB
CODE_TYPE VARCHAR2(5 CHAR)
Initially I tried inserting a few thousand records into both the tables . Then I did some conversion on one column and I populate the result on other column of the same table . So I update the table in bulk mode and I commit for every thousand records .
I have a undo tablespace of 2G with undo_retention=900 , retention guratantee is not set for the undo tablespace .
When I tried the conversion and update on the first table (STAGING_DICT) it took more time for around 2500 records compared to other table and when I increased the number of records to 10000 it threw an error
ORA-30036: unable to extend segment by 8 in undo tablespace 'UNDOTBS'
But I didn't come across this error when I tried the conversion and update on the table for the same 2500 records (STAGING_DICT1) and it was also 10 times faster .
My doubt is does the error ORA-30036 occur because it is saving the undo image of all the four clob columns though I am doing conversion on one column and updating the other column (using only two columns in the update and only one column is affected by update command) ?
Also how is that having less CLOB rows prove more effective by adding one more VARCHAR column which differentiates the code_frament type in the STAGING_DICT1 table than having it as more CLOB columns as in STAGING_DICT table ?
] 
Blog
Search
Help
Members
Register
Login
Home
