MS SQL Server searching functions

Hi,
the Soundex search words that sounds similar.
Does MS SQL Server has some function to make some intuitive search?
For example, for search term database, it should return rows that contains: "database" word, but also rows that contains "Oracle", "MySQL", "MS SQL" etc. terms.

Yes the Soundex string function is implemented in SQL Server and the others like MySQL and Oracle are not close to SQL Server in String functions because SQL Server implemented all. I have posted it a while back check the post below for all the SQL Server String functions. Hope this helps.

http://forums.asp.net/910656/ShowPost.aspx

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or two

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
Select * from database
where NOTES like '%something%'
Is there a way to get results from this query in less then 1-2 second
and how?
Maybe. If possible begin the % with a leading character and try creating a
NC index on the NOTES column. Or might also consider creating a Full-Text
index on the NOTES column and then use CONTAINS or FREETEXT.
HTH
Jerry
"nywebmaster" <scgwebmaster@.yahoo.com> wrote in message
news:1129306144.751022.186010@.o13g2000cwo.googlegr oups.com...
>
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>
|||Thanks.

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or two

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
ID - nvarchar lenth-9
COMPANY - nvarchar lenth-30
NOTES - nvarchar length-250

Select * from database
where NOTES like '%something%'

Is there a way to get results from this query in less then 1-2 second
and how?Hello
I think that using like it would be impossible.
You will get better results when you use full text search (read about it in
books online), however I have not experience with looking for a phrase, but
with single words it works fast.
Alwik

> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ID - nvarchar lenth-9
> COMPANY - nvarchar lenth-30
> NOTES - nvarchar length-250
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?|||On a Athon3200+ (32 bits) home computer
it takes 1692ms to search something like '%RRIDA%' on 393951 rows
table. The maximum length of a row is 3576 bytes. So you only need a
faster CPU and faster memory controler and enough memory to hold the
data pages in memory to achive subsecond time. But IMHO I think this
kind of search is a nonsense for this number of rows.|||So what kind of search are you recommend?|||Am 14 Oct 2005 12:33:10 -0700 schrieb nydefender:

> So what kind of search are you recommend?

What hardware do you use? And how long does it last to get the result? Have
you tried it with an index on NOTES? And i think, a second search should be
much faster then the first one. If you always search on NOTES maybe you can
hold a second table with only PK and field NOTES, which is redundant
(managed by triggers) but can be pinned into memory (DBCC PINTABLE() -
maybe a silly idea, only brainstorming).
Sometimes i have the same problem to find some records out of a big table
where it lasts up to 30 seconds. At first the user knows from
training/docu, that this could need a "long" time to proceed, second i show
a window with a wait-message and something blinking in it, so the user has
not the feeling that the program hangs.

bye,
Helmut|||helmut woess (hw@.iis.at) writes:
> (managed by triggers) but can be pinned into memory (DBCC PINTABLE() -
> maybe a silly idea, only brainstorming).

Yes, DBCC PINTABLE was really a silly idea of Microsoft/Sybase. (Don't
really know who came up with it.) So silly, that in fact in SQL 2005, the
command DBCC PINTABLE is a no-op that performs nothing.

If a table is referenced often enough, it will be in cache anyway, so
PINTABLE has no effect. But if you pin a large table of which only portions
are referenced with some frequency, this means that you are wasting memory
that could have been used for other table, and thus degrade performance.

The only point I can see with PINTABLE is that you have table that you
query so rarely, that it will fall out of the cache. But when you need to
query it, you need the answers snap.

--
Erland Sommarskog, SQL Server MVP, esquel@.sommarskog.se

Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techin.../2000/books.asp|||You want subsecond performance for your query. And the query can return
thousands of rows. How many time the clerk will spend searching for the
correct row?. subsecond querys are needed for routine operations and
they return only the necessary information to do the task, if not, the
worker is wasting his time. When you look for %something%, do you
really know what you are looking for?

In an hospitalizaton patient table, if I look for %seropositive% in the
observations field or even for %positive% I'm pretty sure its for a
report or an adhoc decission suport query and this doesn't need
subsecond response time. SQL Server is an OLTP system, designed for a
lot of small transactions, and this kind of queries is an incorrect use
of the system in my opinion.

You sould use something like Microsoft Search Service or a similar
product.|||Maybe this kind of query is "incorect" but is necessary. Now this query
takes for about 15-20 secs. I try to find a better way. I will try with
full text search.
Thanks to all of you.

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or two

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
Select * from database
where NOTES like '%something%'
Is there a way to get results from this query in less then 1-2 second
and how?
Full Text index this table and run a full population.
The query would look like this
select * from database where contains(NOTES,'something')
Use the wizard to build the FTS index on your table and make sure you run a
full population.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
<scgwebmaster@.yahoo.com> wrote in message
news:1129307133.782947.259880@.f14g2000cwb.googlegr oups.com...
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or two

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
---
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
----
Select * from database
where NOTES like '%something%'
Is there a way to get results from this query in less then 1-2 second
and how?Maybe. If possible begin the % with a leading character and try creating a
NC index on the NOTES column. Or might also consider creating a Full-Text
index on the NOTES column and then use CONTAINS or FREETEXT.
HTH
Jerry
"nywebmaster" <scgwebmaster@.yahoo.com> wrote in message
news:1129306144.751022.186010@.o13g2000cwo.googlegroups.com...
>
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ---
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>|||Thanks.

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or two

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
---
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
----
Select * from database
where NOTES like '%something%'
Is there a way to get results from this query in less then 1-2 second
and how?Maybe. If possible begin the % with a leading character and try creating a
NC index on the NOTES column. Or might also consider creating a Full-Text
index on the NOTES column and then use CONTAINS or FREETEXT.
HTH
Jerry
"nywebmaster" <scgwebmaster@.yahoo.com> wrote in message
news:1129306144.751022.186010@.o13g2000cwo.googlegroups.com...
>
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ---
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>|||Thanks.sql

MS SQL Server 2000 - Search a table with 300,000+ records in less then a second or tw

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
---
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
----
Select * from database
where NOTES like '%something%'
Is there a way to get results from this query in less then 1-2 second
and how?You need to use a Full Text Index to do that kind of search quickly. Look
it up in Books Online.
Adam Machanic
SQL Server MVP
http://www.datamanipulation.net
--
"nywebmaster" <scgwebmaster@.yahoo.com> wrote in message
news:1129306366.597973.143180@.g43g2000cwa.googlegroups.com...
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ---
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>|||Pattern matched searching when the wild card is prefixed to the parameter
cannot use an index and so in most cases, a table scan in employed. If this
is something critical, you might want to look into full text indexing
options.
If you know the pattern upfront, one trick you can use like create a
computed column representing the part of the string and indexing the column.
Anith|||To begin with, insure that NOTES is indexed.
http://www.microsoft.com/technet/pr...s/c0618260.mspx
Performing a LIKE search on '%something%' will not efficeintly utilize an
index on NOTES, however, 'something%' would.
http://msdn.microsoft.com/library/d...dcharacters.asp
If you need to perform fast 'wildcard' type searches, then consider
implemeting Index Server and full-text search. It is a service that runs
along side SQL Server. Just remember that the predicates CONTAINS and
FREETEXT are used for free-text searches, so it will involve making
revisions to some of your queries.
http://msdn.microsoft.com/library/d...r />
_3rqg.asp
"nywebmaster" <scgwebmaster@.yahoo.com> wrote in message
news:1129306366.597973.143180@.g43g2000cwa.googlegroups.com...
> I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ---
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> Is there a way to get results from this query in less then 1-2 second
> and how?
>|||Thanks

MS SQL Server 2000 - Search a table with 250,000+ records in less then a second

I have one table with 300,000 records and 30 columns.
For example columns are ID, COMPANY, PhONE, NOTES ...
----
ID - nvarchar length-9
COMPANY - nvarchar length-30
NOTES - nvarchar length-250
----
Select * from database
where NOTES like '%something%'
----
Is there a way to get results from this query in less then 1-2 second
and how?A few posts below is a remarkably similar question, only the table has
300,000+ rows. I think it applies to your case as well.
<scgwebmaster@.yahoo.com> wrote in message
news:1129305896.814659.100590@.g43g2000cwa.googlegroups.com...
>I have one table with 300,000 records and 30 columns.
> For example columns are ID, COMPANY, PhONE, NOTES ...
> ----
> ID - nvarchar length-9
> COMPANY - nvarchar length-30
> NOTES - nvarchar length-250
> ----
> Select * from database
> where NOTES like '%something%'
> ----
> Is there a way to get results from this query in less then 1-2 second
> and how?
>

MS SQL Server 2000 - Run query as a different user

Can we run some query as a different user working with MS SQL Server 2000?
For example if we connected to the DB in the Enterprise Manager as User1,
can we run some query as User2?
We need to select from a VIEW pulling data from some DB, we can encrypt the
text of this view hiding the target tales and their structure on the target
DB. But since the view works with this DB it should have the permissions to
SELECT from this DB, and that gives the green light to the user to see these
tables in the Enterprise Manager.
It would be nice if we could run the queries inside the view under another
user account and then encrypt the view so that the user will see the view as
it works but should never see the DB tables or the text of the view.
Just D.Of course you can do that, this is one of the benefits of the views.
Example:
Make dbo owner of table employees
Create a view owned by dbo to access some of the fields of employee table
Give permissions to user1 to this view
User1 will be able to access the view but not the employees table
Ben Nevarez, MCDBA, OCP
Database Administrator
"Dmitri Shvetsov" wrote:

> Can we run some query as a different user working with MS SQL Server 2000?
> For example if we connected to the DB in the Enterprise Manager as User1,
> can we run some query as User2?
> We need to select from a VIEW pulling data from some DB, we can encrypt th
e
> text of this view hiding the target tales and their structure on the targe
t
> DB. But since the view works with this DB it should have the permissions t
o
> SELECT from this DB, and that gives the green light to the user to see the
se
> tables in the Enterprise Manager.
> It would be nice if we could run the queries inside the view under another
> user account and then encrypt the view so that the user will see the view
as
> it works but should never see the DB tables or the text of the view.
> Just D.
>
>

MS SQL 2000 and raid controller

I heard that is said that MSQ SQL 2000 doesn't work correctly with raid (for example raid 5)
I have my application in 2 companies. The first one with not a raid controller (even though the computer has a raid controller) executes a backup when no users are online in more than 8 minuter and the file of the database backup is 3,5 GB. The memory of the computer is 2GB.
In the other company when more than 20 users work in the database with a raid controller the backup is occured in 2,6 minutes and the database backup file is more than 6 GB. The memory of this computer is 4 GB.
What is your opinion?
Is this rumor true?
Do you suggest I should apply raid 5 to the first server?
Regards,
ManolisI am not sure what the problem is you think you have identified. Do you want to explain a little more.

If the rumor is true, which I extremely doubt, thousands and thousands of organizations are in trouble. I think you have been talking to sme stuffed shirt wind bag.|||I've used many versions of Microsoft SQL (even before Microsoft bought it). I've used every version since 4.2 (1994) on RAID controllers of various kinds. The only problems I've ever had with RAID controllers have been outright hardware failures, no RAID / MSSQL software related issues.

I would be astonished if there is any problem using SQL Server (any version) with a RAID controller.

-PatP|||I can add one more to Pat's list, but it is narrowly defined. If your raid controller buffers the writes, on a system with heavy i/o, you might get a stale read error when SQL Server reads from the disk again after a write was issued, but held in buffer.

The resolution to that is to decrease the size of the write buffer.

Besides ... without utilizing RAID configurations, have you seen the size of a
petabyte disk drive? :rolleyes: