Inserting instead of updating
Why don't they teach this in school?
During prototyping or early development, I often want to add a column to a table that already has data in it.
The direct approach is alter table, update column.
Once the table gets bigger than a few hundred thousand rows, the update approach gets slow.
The faster approach is to create a new_table as select * from old_table, drop the old_table, and rename the new_table to the old name.
Why is that faster? Over to Tom Kyte.
During prototyping or early development, I often want to add a column to a table that already has data in it.
The direct approach is alter table, update column.
Once the table gets bigger than a few hundred thousand rows, the update approach gets slow.
The faster approach is to create a new_table as select * from old_table, drop the old_table, and rename the new_table to the old name.
Why is that faster? Over to Tom Kyte.
0 Comments:
Post a Comment
<< Home