Writing queries like this is asking for trouble.

Must be created by some automation tool and is extremely difficult for optimizer.

It might be not the index you found, people normally do...

yakko_Warner (10/26/2016)

---

I'm working on a project with a co-worker and I was working on creating a database for some integration of data for his workflow.

I was trying to figure out...

Table "a" is not mentioned in any filter or join condition.

Of course, every its row being applied to every record of the joined recordset of other 2 tables.

Is it how...

1. Normalise your data.

Here I do normalisation on-fly,...

Sterling Ostin (10/24/2016)

---

Jeff - Look at the fourth and sixth small paragraph of my original post. I did exactly what you are suggesting I did not do. I...

I guess this one is self-confessing:

alter table [Order]

add

DateTimeOrdered as dbo.JsonDateTime([Json], 'DateTimeOrdered', 'yyyy-MM-dd HH:mm:ss') persisted

go

create index IX_Order_DateTimeOrdered on [Order](DateTimeOrdered)

What to add?

Probably only this:

Extract all the other values from b..y...

This should be enough:

---------------------[dbo].[Transaction]-----------------

ALTER TABLE [dbo].[Transaction] ADD CONSTRAINT [PK_Transaction] PRIMARY KEY CLUSTERED

(

[TransactionDate] ASC,

[ReferenceNumber] ASC

)

The rest of the parameters, as I can understand, are optional, and adding them to...

Or, if you cannot rely on directionals being the last piece of Address strings, you may wish to use my original join, bit with an additional check that the Suffix...

In this case you can dynamically remove them from Address...

Is there a fixed set of possible directionals?

csj (10/21/2016)

---

It is not a problem to change datatype from length 50 to 100. But it can be a problem to change from 100 to 50 if a row has...

SELECT DATEADD(dd, DATEDIFF(dd, 0, [date/time])/7*7), 0) WeekStart,

, sum(amount1), SUM(amount2)

FROM #Data

GROUP BY DATEPART(wk, DataDate);

No datefirst settings involved.

Converting of week start date to week number I leave to you, as...

Another question: what's gonna happen if I change a column data type from, say, varchar(50) to nvarchar(100)?

Or add a new column in between 2 existing ones?

Jay@Work (10/20/2016)

---

I'm sure it is doable e.g. create a table data type which contains names and values and loop through the values trade=roofer, name=John

I'm just starting to think it...

jeff.born (10/20/2016)

---

Sometimes I just can't see the solutions even when it's right in front of me. Never thought of using a union. or a compound operator.

One question is the...