Is it possible to write a macro statement using Transact-SQL?
Imagine that we have a table named tblA and fields with the almost
same name, for example Field01, Field02,...Field20. (I have named
fields on that way for the better explanation).
Now, suppose that we want to do almost the same update on all of the
fields:
UPDATE tblA
SET Field01 = 100000
UPDATE tblA
SET Field02 = 100000
and so on...
(we must write 20 identical statements). This example is very simple
(please, forget the solution with one statement because it is clear!).
I wrote a simple example because of my next explanation and question.
In some other languages it is not necessarily to write 20 almost
identical statements. I can write something like this:
FOR i: = 1 TO 20
cTemp := CHAR2(i)
REPLACE Field&cTemp with 10000
NEXT i
-- cTemp (using CHAR2 convert function) have a character values: '01',
'02', '03'... etc.
As you can see, with every step through the loop I have changed the
statements using macro Field&cTemp.
Is it possible to write a similar solution in Transact SQL and avoid
20 identical statements?DECLARE @.i int, @.qry varchar(500)
SET @.i=1
WHILE @.i<=20 BEGIN
SET @.Qry='UPDATE tblA SET Field'+Cast(@.i as varchar)+'=100000'
EXEC(@.Qry)
SET @.i=@.i+1
END
This is not the most efficient method but it closely follows your
example(minus the 0 prefix on the first 9 fields). Better would be to build
up the string for a single update to all columns but I'll leave that to you
:)
Mr Tea
"zaratino" <goran.abdic@.zg.htnet.hr> wrote in message
news:b85bv0dq2h7rqf1atoecb7v71cke66aemu@.
4ax.com...
> Is it possible to write a macro statement using Transact-SQL?
> Imagine that we have a table named tblA and fields with the almost
> same name, for example Field01, Field02,...Field20. (I have named
> fields on that way for the better explanation).
> Now, suppose that we want to do almost the same update on all of the
> fields:
> UPDATE tblA
> SET Field01 = 100000
> UPDATE tblA
> SET Field02 = 100000
> and so on...
> (we must write 20 identical statements). This example is very simple
> (please, forget the solution with one statement because it is clear!).
> I wrote a simple example because of my next explanation and question.
> In some other languages it is not necessarily to write 20 almost
> identical statements. I can write something like this:
> FOR i: = 1 TO 20
> cTemp := CHAR2(i)
> REPLACE Field&cTemp with 10000
> NEXT i
> -- cTemp (using CHAR2 convert function) have a character values: '01',
> '02', '03'... etc.
> As you can see, with every step through the loop I have changed the
> statements using macro Field&cTemp.
> Is it possible to write a similar solution in Transact SQL and avoid
> 20 identical statements?|||A couple of questions:
1. Is there a where clause, or is this a single row table?
2. Updating the same row or rows twenty different times is not a very
efficient approach (it will end up taking twenty different log writes!)
3. How are you matching the field with the value?
In general it is far better when it comes to SQL to execute fewer complex
statements than many simpler statements. Building the proper statement and
executing it will be far better. So you could write something like:
--not meant to be compilable, pseudocode only
set @.query = 'UPDATE tblA --hopefully not your real table name'
set @.query = 'SET '
set @.i = 1
while @.i < 20
begin
set @.query = @.query + 'Field' + cast(@.i as varchar(2)) + ' = 100000, '
set @.i = @.i + 1
end
set @.query = @.query + 'WHERE --and your where clause'
exec (@.query)
----
Louis Davidson - drsql@.hotmail.com
SQL Server MVP
Compass Technology Management - www.compass.net
Pro SQL Server 2000 Database Design -
http://www.apress.com/book/bookDisplay.html?bID=266
Note: Please reply to the newsgroups only unless you are interested in
consulting services. All other replies may be ignored :)
"zaratino" <goran.abdic@.zg.htnet.hr> wrote in message
news:b85bv0dq2h7rqf1atoecb7v71cke66aemu@.
4ax.com...
> Is it possible to write a macro statement using Transact-SQL?
> Imagine that we have a table named tblA and fields with the almost
> same name, for example Field01, Field02,...Field20. (I have named
> fields on that way for the better explanation).
> Now, suppose that we want to do almost the same update on all of the
> fields:
> UPDATE tblA
> SET Field01 = 100000
> UPDATE tblA
> SET Field02 = 100000
> and so on...
> (we must write 20 identical statements). This example is very simple
> (please, forget the solution with one statement because it is clear!).
> I wrote a simple example because of my next explanation and question.
> In some other languages it is not necessarily to write 20 almost
> identical statements. I can write something like this:
> FOR i: = 1 TO 20
> cTemp := CHAR2(i)
> REPLACE Field&cTemp with 10000
> NEXT i
> -- cTemp (using CHAR2 convert function) have a character values: '01',
> '02', '03'... etc.
> As you can see, with every step through the loop I have changed the
> statements using macro Field&cTemp.
> Is it possible to write a similar solution in Transact SQL and avoid
> 20 identical statements?|||Thanks a lot Mr Tea and Mr Davidson
It's work (on my more complex task). :-)
And of course - answers:
1. I have a another table (not this one for update). That table
contains circular nodes of hieararchy. This is the reason why I must
first fullfill columns step-by-step.
2. ...it means: yes, I have a WHERE clause and FROM clause (JOIN with
circular table) also.
3. Yes, .log file is written 20 times but I will sucrifise that. I
work with basic data (corporate hierarchy) - not huge set of rows.
4. Of course, matching values in my example is not so simple. I
matching a values with another one 'macro' that read a data form
another table.
...and all of that because you help me! :-)
So,
Thank you once again
Mr Zaratino|||>> Imagine that we have a table named tblA and fields [sic] with the
almost same name, for example Field01, Field02,...Field20. (I have
named fields [sic] on that way for the better explanation). <<
A column is not a field -- nothing like it at all. Since each column
is a separate attribute of the entity in your data model, it would be
VERY unusual to have such a table if you had a proper data model.
However, if I were writing a 1950's file system (files are made of
records which do have fields), then they would probably be a repeating
group -- and a violation of First Normal Form (1NF).
fields [sic]: <<
In SQL an UPDATE works on entire rows (rows are not records), changing
all the columns at the same time.
UPDATE Foobar
SET x = <value1>,
y = <value2>,
z = <value3>,
etc.
If you want to pass the values as parameters, then you can skip some of
them by passing a NULL and having this SET clause in your UPDATE
statement.
SET x = COALESCE (<value1>, x)
Dynamic SQL generation is considered very poor design; it says you have
no data model and no idea what to do until run time.|||>> That table contains circular nodes of hieararchy. This is the reason
why I must
first fullfill columns step-by-step. <<
Do you mean that you are using an adjacency list model for a hierarchy?
If so, look up the nested set model instead. Otherwise, you are not
usingthe power of a set-oriented language and have re-invented a file
system.|||Yes, Celko - everything that you said is correct, I understand UPDATE
statement; sorry for my confusion about 'fields' and 'columns'.
My congratulation, you recognize that I violate 1NF but there is a
good reason for that. I need that look of table for further purpose
(cube). With table like this the next actions are faster...(sometimes
this is even necessarly).
Thanks,
Zaratino
