SQL Server Performance Forum – Threads Archive
JOIN Performance on VLDBHello Does anyone have an idea about the performance issues regarding JOIN’s? If I have two tables UID | INFO
UID | INFO2 and the two of these are 10+ million records, being hit regularly but 10’s of thousands of web users. Every record in the second DB is scanned every night to see what is in INFO2. That is the reason for separation of INFO and INFO2 What kind of impact would I have selecting INFO1 and INFO2 using a JOIN? and Is this silly to think this is possible? Thanks in advance
Your description is a little vague, to be honest. You mention two tables, then you mention two databases. Do you mean that you’re copying the contents of INFO into INFO2 every night? What would your question be … "is it be better to join the two tables on the UID columns and then read INFO2, instead of INFO?" ? If that is your question, then why are you not simply querying the first table?
Piggy-backing on Adriaan, please be more specific. <br />What is the structure of that tables?<br />What are the indexes?<br />What do you want to achieve?<br />What will the queries look like?<br />…<br /><br />Btw, nobody considers even a 10+ rows table a VLDB these days [<img src=’/community/emoticons/emotion-5.gif’ alt=’‘ />]<br /><br />–<br />Frank Kalis<br />SQL Server MVP<br /<a target="_blank" href=http://www.insidesql.de>http://www.insidesql.de</a><br />