Skip to main content

Data purging and Archival from VLDB

How to delete Records from a Very Large Tables ?

If we have to delete a large records based on some conditions like year wise - month wise from very large table and our SLA is without degrading system performance then we will come into problem .We can't use TRUNCATE TABLE and we may run out of disk space and then DELETE query fails and nothing is deleted.


The Real-Problem (from SQL Bible)

A DELETE executes as a transaction where it will not commit until the last record is deleted. DELETE physically removes rows one at a time and records each deleted row in the transaction log. If the number of records in a table is small, a straight out DELETE is fine, however, with a large table the DELETE will cause the transaction log to grow. When this happens the system IO performance is degraded. Also, the table will be locked which will affect the application. The person executing the DELETE query will usually panic because what seems to be a simple task takes a long time and there is no visible progress from Query Analyzer. If the transaction log runs out of disk space then the DELETE fails and the transaction rolls back and nothing ends up deleted.


A Solution:
A solution is to use the combination of a WHILE loop and @@ROWCOUNT to delete the records in manageable blocks
Below is a script that you can use to do bulk deletions.

select count(*) from prps_delete where left(erdat,4)='2012'

SET ROWCOUNT 10000

WHILE 1 = 1
      BEGIN
      
      -------------------------------------------------
      DELETE prps_delete WHERE left(erdat,4)='2012'
     
      IF @@ROWCOUNT = 0
            BREAK
      END
SET ROWCOUNT 0

(10000 row(s) affected)

(10000 row(s) affected)

(10000 row(s) affected)

(10000 row(s) affected)

.
.
.
Cont………….

(526 row(s) affected)
select count(*) from prps_delete where left(erdat,4)='2012'


Thanks
SumanJha_Accidental_DBA

Comments

Popular posts from this blog

How to encrypt and decrypt Table data in postgres

For encrypting and decrypting , we must use the bytea data type on the column which we implement. Bcoz bytea will use the pgcrypto method by default. However, you will need to create the pgcrypto extension to enable these functions as they are not pre-defined in PostgreSQL/PPAS. Example CREATE EXTENSION pgcrypto; CREATE TABLE userinfo (username varchar(20), password bytea); >>    Inserting the data in an encrypted format INSERT INTO userinfo VALUES(' suman ',encrypt('111222','password','aes')); select * from userinfo ; >>    Retrieving the data as decrypted format SELECT decrypt(password,decode('password','escape'::text),'aes'::text) FROM userinfo; Thanks for reading Plz dont forget to like Facebook Page.. https://www.facebook.com/pages/Sql-DBAcoin/523110684456757

How to recover msdb database from suspect mode

 It was Monday 9 th Jun 47 degr. temperature of Delhi-NCR. Temperature was like boiling me and database. When I reached my office( @ 8.45 am) got an alert from one of Server. “MSDB is in suspected mode” At the same time comes in my mind, this issue will boil me today.. I just tried to cool my self through cold drink then connected server from my local system using windows authentication mode..

History of MySQL from AB Corp to Cloud Database

MySQL was created by a Swedish company, MySQL AB, founded by David Axmark, Allan Larsson and Michael "Monty" Widenius. Original development of MySQL by Widenius and Axmark began in 1994. The first version of MySQL appeared on 23 May 1995. Its name is a combination of "My", the name of co-founder Michael Widenius's daughter,and "SQL", the abbreviation for Structured Query Language. ·          23 May 1995 - First internal release ·          Year 1996 - Version 3 o     Simple CRUD operations o     January 1997 Windows version was released on 8 January 1998 for Windows 95 and NT o     production release 1998, from www.mysql.com ·          Year 2002 - Version 4 o     MyISAM o     unions o     Tracking o     B-trees o     s...