Rob Farley

Rob Rob Farley has been consulting in IT since completing a Computer Science degree with first class honours in 1997. Before moving to Adelaide, he worked in consultancies in Melbourne and London. He runs the development department in one of Australia's leading IT firms, as well as doing database application consultancy and training. He heads up the Adelaide SQL Server User Group, and holds several Microsoft certifications.

Rob has been involved with Microsoft technologies for most of his career, but has also done significant work with Oracle and Unix systems. His preferred database is SQL Server and his preferred language is C#. Recently he has been involved with Microsoft Learning in the US, creating and reviewing new content for the next generation of Microsoft exams.

Over the years, Rob's clients have included BP Oil, OneLink Transit, Accenture, Avanade, Australian Electorial Commission, the Chartered Institute of Personnel and Development, the Royal Borough of Kingston, Help The Aged, Unisys, Department of Treasury and Finance (Vic), National Mutual, the Bible Society and others.

Did you mean to come here? My blog is now at

09 November 2005

Number padding in T-SQL

I saw a SQL Tip on padding numbers, which seemed really lousy.

The way I pad numbers is to use 'right', rather than 'left' and 'len'. And certainly not 'case'.

They wrote:

SELECT value,LEFT('000000',(6 -LEN(value ))) + CONVERT(VARCHAR,value) AS Formattedvalue
FROM #ValueTable

And I'm suggesting:

SELECT value,RIGHT('000000' + CONVERT(VARCHAR,value), 6) AS Formattedvalue
FROM #ValueTable

The way I see it - if you have something you want to appear on the RIGHT, with some unknown amount of stuff to the left, then it makes sense to use RIGHT. If I don't want it to break when the length of the value is too long, then a case statement will do the job.