Comments on: Why You Should Never Use SELECT Star (*)

By: Curious George

Curious George — Mon, 30 Jul 2007 19:48:33 +0000

I agree with the SELECT * fiasco. But I wonder what your feelings are on using COUNT(*) to pull a quick return of total record count. Is it viewed as wasteful, also?

EG: SELECT COUNT(*) FROM MY_TABLE

I kinda abuse that sql statement a lot, especially when I embed sql in VB code, to get a quick recordcount of a table so I can iterate through it w/o having to open it as a dynamic ADO so I can use the ADO.Recordset.Recordcount function. (IE: I find it wasteful to open a recordset as a larger, dynamic object just so I can have access to it’s recordcount, so I jury rig around it by using a COUNT(*) of the table first.)

Just curious about your thoughts on this.

By: jlango

jlango — Fri, 26 Oct 2007 01:17:53 +0000

COUNT(*) is one of those damned-if-ya-do situations.
Usually, the only way that you can get an accurate record count is to use COUNT(*). So, you’re kind of stuck using it.

I haven’t written and VB code in a few years now. I need to look at some of old VB code before I can speak to your question.

By: [re-arrange]

[re-arrange] — Tue, 23 Nov 2010 20:32:51 +0000

I don’t view count(*) as wasteful, since the resultSet of total record count doesn’t really return a large resultSet. Usually in most programming language it returns one single number.

Most database can handle the row count using that count(*), btw.

By: hcpcs code books

hcpcs code books — Sun, 12 Dec 2010 17:29:29 +0000

Hello admin! This is very helpful posting thanks.

By: ssokolow

ssokolow — Fri, 17 Dec 2010 17:19:40 +0000

I forget which specific ones (It’s been a while) but some databases actually apply extra optimizations on COUNT(*), relying on the implicit “Count returned rows” meaning.

I could be wrong, but I think it has something to do with how COUNT(column) interacts with NULL… possibly in the context of indexing.

By: Hcpcs Code Lookup

Hcpcs Code Lookup — Wed, 23 Mar 2011 15:38:41 +0000

Hello admin, thanks for posting and link back to my site.

By: Hcpcs Codes

Hcpcs Codes — Thu, 24 Mar 2011 04:52:30 +0000

Hello admin, thanks for posting and link back to my site.

By: Ricky

Ricky — Fri, 15 Apr 2011 09:04:17 +0000

“When you run a SELECT *, most databases will cache which fields are returned from this. This means that if you add a field to a table, an existing SELECT * query will not return the field!.”

This is certainly bullshit.

By: Dave D

Dave D — Thu, 26 May 2011 10:35:18 +0000

Select * is wrong because it broke your app ? Sounds more like your app didn’t handle a result that the query could have returned. I use Select * in some code (& Select field1, field2 etc in other code) because I want all fields & my code is robust enough to handle whatever result the query returns. Also what Ricky says about BS is correct.

By: jlango

jlango — Thu, 26 May 2011 11:18:20 +0000

Okay, I wrote this like 5 years ago. I can guarantee that this behavior happened then. I haven’t tested it since then, and using ORM probably makes this moot anyway. But since you’re all having so much fun commenting on a 5-year old post, go for it. After, you and Ricky can go find a post from 2003 complaining that T-SQL doesn’t have try/catch statements, and call that bullshit too.