What if you want to have one data series use one subset of your data, while another uses a different subset?

Steve

(aka sgmunson)

:-):-):-)

Just an FYI, my test took place over a year and a half ago now, as did the original post on this topic, and while I was certainly pleased with...

Wayne,

My WITH clause that precedes the function calls had to be prefixed to your code, and as yours has to do ALL the extra comparisons, it loses on performance by...

Ok, that's a pretty decent idea, but then I have to see which group level is common across the group, which amounts to the same problem I have to begin...

It's like this - pretend your data is like this:

GRP_FLD PATH_FLD

AABBCCD \\server1\share1\apath1\abc

AABBCCD \\server1\share1\apath1\def

AABBCCD \\server1\share1\apath1\ghi

EEFFGGH \\server2\share2\apath1\abc

EEFFGGH \\server2\share2\bpath2\abc

What is needed is for each value in the GRP_FLD, determine the characters in...

You may also find that it depends on exactly how the import is taking place... I once had to work with data from an Informix database on a Linux...

Thanks - I'll check it out...

Steve

(aka sgmunson)

:-):-):-)

Yes, and besides, there's just no way even a neural net can handle names like Terry, Chris, Pat (in this case, maybe even an adrogenous individual (SNL reference, LOL)), Jamie,...

The important thing to remember is that being able to update a record requires that there be a unique record or set of records to update, such that there's no...

Just to let everyone know, Chris' solution works quite nicely. The key is in realizing that recursive CTE's recursive references only retrieve the data from the last iteration,...

Perhaps I misunderstand how this is supposed to work... I would have thought that the objective of allowing recursion is precisely to gather a series of records, and...

Interesting... I tried the "simple repro" query that Steve Kass submitted and sure enough, row_number() does not get properly evaluated - it's as if it gets a fresh evaluation...

Why? That's not the path this is going to take... The level is only going to be relative to the root level, not to any other value. ...

I'd really love to understand the reasoning that went into that post's comment that "the result is not the expected one". Here's what SQL Server 2005 Developer Edition...

That gave me at least an idea on how to use the windowed COUNT aggregate. Here's my latest attempt, and it's results. I don't quite understand...