Home/ Questions/Q 1041629
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-16T15:22:57+00:00

I have a table that has about 400,000+ rows. I am writing some pattern

  • 0

I have a table that has about 400,000+ rows. I am writing some pattern matching code but need to clean up a column before I do so. This boils down to doing a replace-like operation.

I tried listing them all out one at a time…

 Update T_ADDRESS set ADDR_LINEONE = REPLACE(ADDR_LINEONE,' southeast ',' se ')
 Update T_ADDRESS set ADDR_LINEONE = REPLACE(ADDR_LINEONE,' southwest ',' sw ')

Since I have over 500 of these…it took too long.

Now I am trying to nest them…

 Update T_ADDRESS set ADDR_LINEONE = REPLACE(REPLACE(ADDR_LINEONE,' southwest ',' sw '),' southeast ',' se ')

But this is still painfully slow. I need to make this code work on tables of all sizes (1 record to 5 million records).

Anyone have any advice? I am using SQL Server by the way.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    You have to always scan the table end-to-end no matter how fancy you do the REPLACE. This is what is killing performance, and it cannot be changed since you have to way of indexing the ADDR_LINEONE field in any sensible manner.

    Since this should be a one-time only operation, the long time should not matter.

    If this is a repeated operation, then your problem is not here, is in how you load the data into the table: do the transformation before you save the data, otherwise you stand no chance.

    • 0