@jeff Moden:

Now, my question is, what is the actual functionality that the OP is asking for in the original post? Although it looks like “using VALUES” in a...

Seems legit, thanks 🙂 But the syntax, man... I'll probably just write my own set of methods...

create function EqualI2(@i1 int, @i2 int)...

create function EqualI3(@i1 int, @i2 int, @i3 int)...

...

create function...

@cadavre: The function should be able to compare multiple values (for example up to 20), then I would need to write a lot of 'default's to compare just...

Yes, with stored procedures this is easily achievable, but I cannot use it in scenarios like:

declare @t table (Id int, I1 int, I2 int, I3 int, I4 int, I5 int)

--...

Thanks for reply, willl try to find some ProgressDB forum and ask there.

Beautiful, thanks 🙂

@Phil - thanks, much simpler than my solution 🙂

Got it 🙂

First we need a table function which returns all integers between x and y (the function is stolen from somewhere, don't remember where...):

create function [dbo].[f_table_numbers] (@st bigint, @end...

Hm... I could create table valued function which gets an id from first table and the number of repeats from second table and performs a cross join... Will try it...

I want my results to be grouped, your query is totally correct. I was just curious if there is any reason behind the decision, that you used dense_rank() instead of...

It is exactly the same as yours, I just replaced dense_rank function with row_number:

SELECT Username,

DateFrom = MIN([Date]),

DateTo = MAX([Date]),

CodeFrom = MIN(Code),

CodeTo = MAX(Code)

FROM (

SELECT Username, [Date], Code,

rn...

@ChrisM@Work - that's it, thanks 🙂

Is there any specific reason why you used dense_rank function? I tried with two row_number functions and it still works ok.

@ColdCoffee: no, I tried that. The query

select

row_number() over (order by Code) RowNumber,

row_number() over (partition by username order by code) RowNumberWithPartition,

row_number() over (order by Code) - row_number()...

Wow, how could I miss this one 🙂 Thanks, that's exactly what I needed.

One question remains - what is the best way to somehow save the settings of the current...

Interesting, I got completely opposite times. Will test it again though, maybe I forgot clustered index on ShiftsTable...

I am using nolock because the inserts should not be blocked in any...