Don't think so.. Couldn't get any 'shorthand' to work anyway...

OTOH, I don't see that the 'correct' syntax should look so overwhelmingly long...?

alter table test alter column col1 datetime null

alter table...

Why do you want to do this, and what should the user be told? ...and if told, what should the user then do?

Being put on hold due to locking, isn't...

What is the datatype of the column MonthReported?

Is it actually a char that has literal values like '09/01/06' or is it a datetime and '09/01/06' only happens to be the...

The wildcard works just the same with LIKE as PATINDEX (you can see the example above)

For ranges, you just specify the allowed/disallowed chars you want to filter on one after...

It does seem like LIKE and a wildcard could solve it...?

This looks like it's working anyway..

create table #x ( a varchar(20) not null )

insert #x

select 'abcde' union...

Do you know how that algorithm works?

(as I have no clue..)

/Kenneth

For starters... It seems like you don't want all data in your table in the file...?

If that is the case, what do you want to use to select the rows...

...and how is just 'stored procedure' the 'solution'...?

The solution to what?

It would probably be easiest if you could post some samples of your data. That way we can see what can or cannot be done with it.

/Kenneth

It's nothing very complicated about it.

DISTINCT and GROUP BY are just two ways of essentially producing the same result - duplicates are removed.

Since your query produces many duplicates in the...

select t1.field1, t2.field3, t2.field4

from table1 t1

 inner join table2 on t2.field3 = t1.field1

where t1.field1 = 'char 200' 

(gives you many dupe rows)

or.....

select DISTINCT t1.field1, t2.field3, t2.field4

from table1 t1

 inner join table2 on t2.field3...

You can use a command promt and bcp.

bcp "select * from myTable where foo between 35 million and 70 million" queryout c:\myfile.txt -c -t; -T

(you get the picture)

There's more info...

Mmm.. slippery bastards those rounding errors..

It does indeed seem like your case variant is the most straightforward and accurate way.

/Kenneth

If I read your question correctly, what you're really asking is a modeling question.

There are tons and tons of material written on datamodeling, and there is no simple answer that...

Indeed it does. Interesting, haven't noticed that before.

Though it appears that if we replace FLOOR/CEILING with ROUND, it will take care of that. It seems to work the same as the...