Very good point!

I don’t know of any. (Though i am not an expert.)

It seems that every field would need to be scanned, resulting in a full table scan. Indexing won’t even help because it would be indexing off the description and not the part that you want.

Thank you for the suggestion which was obviously in my mind. I am new into the team and company, and the application has been built and stabilised. I need to study it better before i make any changes to the DB.
But at this time, I have been asked to get a quick report where I need to filter on something like this. Is there any other quick solution running in your mind?

Ideally there should be a type field instead of searching a name for the type.

If there is a list that will be managed daily, perhaps another TABLE would be best. Then the query could use a join and not have to change.

SELECT DISTINCT Building WHERE Building IS NOT NULL
AND Building NOT LIKE ‘%School%’
AND Building NOT LIKE ‘%Church%’
AND Building NOT LIKE ‘%Rectory%’
AND Building NOT LIKE ‘%Public House%’
AND Building NOT LIKE ‘%Retail Park%’
AND Building NOT LIKE ‘%Railway Station%’
AND Building NOT LIKE ‘%Post Office%’
AND Building NOT LIKE ‘%Court%’
AND Building NOT LIKE ‘%Floor%’
AND Building NOT LIKE ‘%Flat%’
AND Building NOT LIKE ‘%Former%’
AND Building NOT LIKE ‘%Lodge%’

basically as per my requirement, the list in the not like might increase daily, I dont find the query good with the whole list as above. Is there a way I can avoid it.

Please suggest.

Your first select statement is not valid. I am not sure if it could ever execute, because you are not specific about what to be search for NOT IN.

Anyways below is a working example, consider this example and let us know if this is what you were expecting.

Example Script :

create table #die_domain_log ( Value int identity , source_flag int , target_flag int, log_id int )

insert into #die_domain_log values ( 1, NULL , 103)
insert into #die_domain_log values ( 1, NULL , 103)
insert into #die_domain_log values ( 1, NULL , 103)
insert into #die_domain_log values ( 1, NULL , 103)
insert into #die_domain_log values ( 1, NULL , 103)

insert into #die_domain_log values ( NULL , 1, 103)
insert into #die_domain_log values ( NULL , 1, 103)
insert into #die_domain_log values ( NULL , 1, 103)
insert into #die_domain_log values ( NULL , 1, 103)
insert into #die_domain_log values ( NULL , 1, 103)

Select * from #die_domain_log

– Here you need to be specific what to search for NOT IN, I have include Value as a search criteria for NOT IN.

SELECT value
FROM #die_domain_log
WHERE source_flag =1
AND log_id=103
AND value
NOT IN
(
SELECT i.value
FROM #die_domain_log i
WHERE i.target_flag =1
AND i.log_id=103
)

– OR ( below is same as you posted)

SELECT value FROM #die_domain_log WHERE source_flag =1 AND log_id=103
EXCEPT
SELECT i.value FROM #die_domain_log i WHERE i.target_flag =1 AND i.log_id=103

~ IM.

What i came across is this
i have a table with around 90000 rows .
45000 where source_flag=1 and 45000 where target_flag=1

i was using a not in query like this

SELECT value FROM die_domain_log WHERE source_flag =1 AND log_id=103
NOT IN ( SELECT i.value FROM die_domain_log i WHERE i.target_flag =1 AND i.log_id=103 )

When executed it didnt returned any results for like 15-20 mins i stoped it .

Then i changed it to except it took less than a sec to return the result .

SELECT value FROM die_domain_log WHERE source_flag =1 AND log_id=103
EXCEPT
SELECT i.value FROM die_domain_log i WHERE i.target_flag =1 AND i.log_id=103

why is this happening ?

I need Result Drop In To New Table

You can also use red-gate to compare data and sql it’s very good DB tool.

Now,how do i compare both the tables,such that.when compared with original database if any record or column is updated in the local database,it should return me updated records.

Using Java and SQL ,how can i do this.Please reply as earle as possible.

Thanks in Advance!!!

It helped me lot in my problem.Thanks lot.

Thank you very much.
See you again.

Nandkishor

I am sure there must be other ways to do this, I used a cursor and while loop,

This can also be done with one cursor with out using a while loop, since the third column in first table contains duplicate rows, the cursor will be executed many times ( how many duplicate rows you have in column 3 in first table, that many times) I thought that would unneccarily consume performance so I chose cursor and while loop to increase performance of the query.

In short, this is also performence consuming, but I could not think of anything else,.. this is atleast a better one in my view..

Here starts your code….

CREATE TABLE TABLE_1 (TSID INT, TSDATE VARCHAR(10), ENAME VARCHAR(10))

INSERT INTO TABLE_1 (TSID , TSDATE, ENAME )
SELECT 100, ‘7-21-2008′ ,’NAME1′UNION ALL
SELECT 101, ‘7-21-2008′, ‘NAME1′UNION ALL
SELECT 102, ‘7-22-2008′, ‘NAME1′UNION ALL
SELECT 103, ‘7-26-2008′, ‘NAME1′UNION ALL
SELECT 104, ‘7-23-2008′, ‘NAME2′UNION ALL
SELECT 105,’7-25-2008′, ‘NAME2′

CREATE TABLE TABLE_2 (NODATES VARCHAR(10))

INSERT INTO TABLE_2 (NODATES)
SELECT ‘7-21-2008′UNION ALL
SELECT ‘7-22-2008′ UNION ALL
SELECT ‘7-23-2008′UNION ALL
SELECT ‘7-24-2008′UNION ALL
SELECT ‘7-25-2008′UNION ALL
SELECT ‘7-26-2008′

We created two tables, and inserted the sample data given by you,

I am creating two temporary tables, and then I dropped those two temporary tables at the end of the code.

SET NOCOUNT ON
DECLARE @ENAME VARCHAR(10)
DECLARE @DATE VARCHAR(10)
DECLARE @VAR INT
CREATE TABLE #RESULT ( DATE VARCHAR(10), NAME VARCHAR(10))
CREATE TABLE #TEMP1 (ID INT IDENTITY, ENAME VARCHAR(10))
INSERT INTO #TEMP1(ENAME) SELECT DISTINCT ENAME FROM TABLE_1
DECLARE @CUR CURSOR
SET @CUR = CURSOR FOR
SELECT DISTINCT NODATES FROM TABLE_2
OPEN @CUR
FETCH NEXT
FROM @CUR INTO @DATE
WHILE @@FETCH_STATUS = 0
BEGIN
SET @VAR = 1
WHILE @VAR < = (SELECT COUNT(1) FROM #TEMP1)
BEGIN
SELECT @ENAME = ENAME FROM #TEMP1 WHERE ID = @VAR
IF EXISTS ( SELECT TSDATE FROM TABLE_1 WHERE TSDATE = @DATE)
BEGIN
IF EXISTS (SELECT ENAME FROM TABLE_1 WHERE TSDATE = @DATE AND ENAME @ENAME )

BEGIN
INSERT INTO #RESULT (DATE , NAME ) SELECT @DATE ,@ENAME
END
END
ELSE INSERT INTO #RESULT (DATE , NAME) SELECT @DATE, @ENAME
SET @VAR = @VAR +1
END
FETCH NEXT
FROM @CUR INTO @DATE
END
CLOSE @CUR
DEALLOCATE @CUR
SELECT * FROM #RESULT ORDER BY NAME , DATE
DROP TABLE #TEMP1
DROP TABLE #RESULT
GO

Result :

7-23-2008 Name1
7-24-2008 Name1
7-25-2008 Name1
7-21-2008 Name2
7-22-2008 Name2
7-24-2008 Name2
7-26-2008 Name2

Hope this helps,
Thanks,
Imran.

I have to compare this 2 tables…,

Now which dates are not in table1 that date and that ename have to come in select statement

Table1:

TSID TSdate EName
100 7-21-2008 Name1
101 7-21-2008 Name1
102 7-22-2008 Name1
103 7-26-2008 Name1
104 7-23-2008 Name2
105 7-25-2008 Name2

Table2:

Nodates
7-21-2008
7-22-2008
7-23-2008
7-24-2008
7-25-2008
7-26-2008

output:
Date Name
7-23-2008 Name1
7-24-2008 Name1
7-25-2008 Name1
7-21-2008 Name2
7-22-2008 Name2
7-24-2008 Name2
7-26-2008 Name2

how to do that