Searching for multiple keywords using SQL Server stored procedure - sql-server-2008

I'm going to search my database (SQL Server 2008) using a stored procedure. My users can enter keyword(s) in a textbox (keywords can be separated using , for instance).
Currently I'm using something like this:
keyword like N"%'+#SearchQuery%'%"
(keyword is a nvarchar column in my table, and #SearchQuery is the input to my stored procedure)
It works fine but what if user types several keywords: apple,orange, banana
Should I limit number of my keywords? How should I write my stored procedure if I have more than one keyword? How should I pass my user input to the stored procedure? I should pass apple, orange, banana as a whole phrase and then I should parse them in my stored procedure, or I should separate my keywords and send 3 keywords? How can I query these 3 keywords? A for loop?
What are best practices for performing such queries?
thanks

Do the parsing of the keywords in your application. SQL is not the best place for string manipulation.
Send the keywords as a table valued parameter (ie : http://www.mssqltips.com/sqlservertip/2112/table-value-parameters-in-sql-server-2008-and-net-c/ ) then you aren't limited to a fixed number of keywords.
Add the wildcards to the parameter in the stored procedure
update #keywords set keyword = '%'+keyword+'%'
filter your results by joining your source data to this table
eg:
SELECT result
FROM source
INNER JOIN #keywords keywords
ON source.keyword LIKE keywords.keyword

It depends on:
* How big it's your database.
* How often users will search for something.
* How precise results users except.
LIKE is not performance daemon, especially starting with %.
Maybe you should try full search text?
If you would like stay with LIKE (it will works only for small tables) I would try something like:
Split intput by , character (insert them into table as podiluska suggested is a good idea).
Build query for each token and UNION all results. Or run it in loop for each token and insert results to temporary table.
If you need some precise results (i.e. only records matches all 3 words) you can select most matching results from temporary results built above.

You could use CTE to split the string of keywords in a temporary table and then use it as you like. The keyword list can even have numbers or any characters, like %$<> or what you want, just remember comma is the string separator
DECLARE #CommaSeparatorString VARCHAR(MAX),
#CommaSeparatorXML XML
DECLARE #handle INT
SELECT #CommaSeparatorString = 'apple,orange,banana'
SELECT #CommaSeparatorString = REPLACE(REPLACE(#CommaSeparatorString,'<','$^%'),'>','%^$')
SELECT #CommaSeparatorXML = CAST('<ROOT><i>' + REPLACE(#CommaSeparatorString, ',', '</i><i>') + '</i></ROOT>' AS XML)
SELECT REPLACE(REPLACE(c.value('.', 'VARCHAR(100)'),'$^%','<'),'%^$','>') AS ID
FROM (SELECT #CommaSeparatorXML AS CommaXML) a
CROSS APPLY CommaXML.nodes('//i') x(c)
Result:
ID
------
apple
orange
banana

Related

how to pass multiple variables in WHERE ... IN in stored procedure? [duplicate]

I have a column in one of my table where I store multiple ids seperated by comma's.
Is there a way in which I can use this column's value in the "IN" clause of a query.
The column(city) has values like 6,7,8,16,21,2
I need to use as
select * from table where e_ID in (Select city from locations where e_Id=?)
I am satisfied with Crozin's answer, but I am open to suggestions, views and options.
Feel free to share your views.
Building on the FIND_IN_SET() example from #Jeremy Smith, you can do it with a join so you don't have to run a subquery.
SELECT * FROM table t
JOIN locations l ON FIND_IN_SET(t.e_ID, l.city) > 0
WHERE l.e_ID = ?
This is known to perform very poorly, since it has to do table-scans, evaluating the FIND_IN_SET() function for every combination of rows in table and locations. It cannot make use of an index, and there's no way to improve it.
I know you said you are trying to make the best of a bad database design, but you must understand just how drastically bad this is.
Explanation: Suppose I were to ask you to look up everyone in a telephone book whose first, middle, or last initial is "J." There's no way the sorted order of the book helps in this case, since you have to scan every single page anyway.
The LIKE solution given by #fthiella has a similar problem with regards to performance. It cannot be indexed.
Also see my answer to Is storing a delimited list in a database column really that bad? for other pitfalls of this way of storing denormalized data.
If you can create a supplementary table to store an index, you can map the locations to each entry in the city list:
CREATE TABLE location2city (
location INT,
city INT,
PRIMARY KEY (location, city)
);
Assuming you have a lookup table for all possible cities (not just those mentioned in the table) you can bear the inefficiency one time to produce the mapping:
INSERT INTO location2city (location, city)
SELECT l.e_ID, c.e_ID FROM cities c JOIN locations l
ON FIND_IN_SET(c.e_ID, l.city) > 0;
Now you can run a much more efficient query to find entries in your table:
SELECT * FROM location2city l
JOIN table t ON t.e_ID = l.city
WHERE l.e_ID = ?;
This can make use of an index. Now you just need to take care that any INSERT/UPDATE/DELETE of rows in locations also inserts the corresponding mapping rows in location2city.
From MySQL's point of view you're not storing multiple ids separated by comma - you're storing a text value, which has the exact same meaing as "Hello World" or "I like cakes!" - i.e. it doesn't have any meaing.
What you have to do is to create a separated table that will link two objects from the database together. Read more about many-to-many or one-to-many (depending on your requirements) relationships in SQL-based databases.
Rather than use IN on your query, use FIND_IN_SET (docs):
SELECT * FROM table
WHERE 0 < FIND_IN_SET(e_ID, (
SELECT city FROM locations WHERE e_ID=?))
The usual caveats about first form normalization apply (the database shouldn't store multiple values in a single column), but if you're stuck with it, then the above statement should help.
This does not use IN clause, but it should do what you need:
Select *
from table
where
CONCAT(',', (Select city from locations where e_Id=?), ',')
LIKE
CONCAT('%,', e_ID, ',%')
but you have to make sure that e_ID does not contain any commas or any jolly character.
e.g.
CONCAT(',', '6,7,8,16,21,2', ',') returns ',6,7,8,16,21,2,'
e_ID=1 --> ',6,7,8,16,21,2,' LIKE '%,1,%' ? FALSE
e_ID=6 --> ',6,7,8,16,21,2,' LIKE '%,6,%' ? TRUE
e_ID=21 --> ',6,7,8,16,21,2,' LIKE '%,21,%' ? TRUE
e_ID=2 --> ',6,7,8,16,21,2,' LIKE '%,2,%' ? TRUE
e_ID=3 --> ',6,7,8,16,21,2,' LIKE '%,3,%' ? FALSE
etc.
Don't know if this is what you want to accomplish. With MySQL there is feature to concatenate values from a group GROUP_CONCAT
You can try something like this:
select * from table where e_ID in (Select GROUP_CONCAT(city SEPARATOR ',') from locations where e_Id=?)
this one in for oracle ..here string concatenation is done by wm_concat
select * from table where e_ID in (Select wm_concat(city) from locations where e_Id=?)
yes i agree with raheel shan .. in order put this "in" clause we need to make that column into row below code one do that job.
select * from table where to_char(e_ID)
in (
select substr(city,instr(city,',',1,rownum)+1,instr(city,',',1,rownum+1)-instr(city,',',1,rownum)-1) from
(
select ','||WM_CONCAT(city)||',' city,length(WM_CONCAT(city))-length(replace(WM_CONCAT(city),','))+1 CNT from locations where e_Id=? ) TST
,ALL_OBJECTS OBJ where TST.CNT>=rownum
) ;
you should use
FIND_IN_SET Returns position of value in string of comma-separated values
mysql> SELECT FIND_IN_SET('b','a,b,c,d');
-> 2
You need to "SPLIT" the city column values. It will be like:
SELECT *
FROM table
WHERE e_ID IN (SELECT TO_NUMBER(
SPLIT_STR(city /*string*/
, ',' /*delimiter*/
, 1 /*start_position*/
)
)
FROM locations);
You can read more about the MySQL split_str function here: http://blog.fedecarg.com/2009/02/22/mysql-split-string-function/
Also, I have used the TO_NUMBER function of Oracle here. Please replace it with a proper MySQL function.
IN takes rows so taking comma seperated column for search will not do what you want but if you provide data like this ('1','2','3') this will work but you can not save data like this in your field whatever you insert in the column it will take the whole thing as a string.
You can create a prepared statement dynamically like this
set #sql = concat('select * from city where city_id in (',
(select cities from location where location_id = 3),
')');
prepare in_stmt from #sql;
execute in_stmt;
deallocate prepare in_stmt;
Ref: Use a comma-separated string in an IN () in MySQL
Recently I faced the same problem and this is how I resolved it.
It worked for me, hope this is what you were looking for.
select * from table_name t where (select (CONCAT(',',(Select city from locations l where l.e_Id=?),',')) as city_string) LIKE CONCAT('%,',t.e_ID,',%');
Example: It will look like this
select * from table_name t where ',6,7,8,16,21,2,' LIKE '%,2,%';

Unable to find Comma Separated Values from Table in MySQL [duplicate]

I have a table say, ITEM, in MySQL that stores data as follows:
ID FEATURES
--------------------
1 AB,CD,EF,XY
2 PQ,AC,A3,B3
3 AB,CDE
4 AB1,BC3
--------------------
As an input, I will get a CSV string, something like "AB,PQ". I want to get the records that contain AB or PQ. I realized that we've to write a MySQL function to achieve this. So, if we have this magical function MATCH_ANY defined in MySQL that does this, I would then simply execute an SQL as follows:
select * from ITEM where MATCH_ANY(FEAURES, "AB,PQ") = 0
The above query would return the records 1, 2 and 3.
But I'm running into all sorts of problems while implementing this function as I realized that MySQL doesn't support arrays and there's no simple way to split strings based on a delimiter.
Remodeling the table is the last option for me as it involves lot of issues.
I might also want to execute queries containing multiple MATCH_ANY functions such as:
select * from ITEM where MATCH_ANY(FEATURES, "AB,PQ") = 0 and MATCH_ANY(FEATURES, "CDE")
In the above case, we would get an intersection of records (1, 2, 3) and (3) which would be just 3.
Any help is deeply appreciated.
Thanks
First of all, the database should of course not contain comma separated values, but you are hopefully aware of this already. If the table was normalised, you could easily get the items using a query like:
select distinct i.Itemid
from Item i
inner join ItemFeature f on f.ItemId = i.ItemId
where f.Feature in ('AB', 'PQ')
You can match the strings in the comma separated values, but it's not very efficient:
select Id
from Item
where
instr(concat(',', Features, ','), ',AB,') <> 0 or
instr(concat(',', Features, ','), ',PQ,') <> 0
For all you REGEXP lovers out there, I thought I would add this as a solution:
SELECT * FROM ITEM WHERE FEATURES REGEXP '[[:<:]]AB|PQ[[:>:]]';
and for case sensitivity:
SELECT * FROM ITEM WHERE FEATURES REGEXP BINARY '[[:<:]]AB|PQ[[:>:]]';
For the second query:
SELECT * FROM ITEM WHERE FEATURES REGEXP '[[:<:]]AB|PQ[[:>:]]' AND FEATURES REGEXP '[[:<:]]CDE[[:>:]];
Cheers!
select *
from ITEM where
where CONCAT(',',FEAURES,',') LIKE '%,AB,%'
or CONCAT(',',FEAURES,',') LIKE '%,PQ,%'
or create a custom function to do your MATCH_ANY
Alternatively, consider using RLIKE()
select *
from ITEM
where ','+FEATURES+',' RLIKE ',AB,|,PQ,';
Just a thought:
Does it have to be done in SQL? This is the kind of thing you might normally expect to write in PHP or Python or whatever language you're using to interface with the database.
This approach means you can build your query string using whatever complex logic you need and then just submit a vanilla SQL query, rather than trying to build a procedure in SQL.
Ben

SQL - Query to find if a string contains part of the value in Column

I am trying to write a Query to find if a string contains part of the value in Column (Not to confuse with the query to find if a column contains part of a string).
Say for example I have a column in a table with values
ABC,XYZ
If I give search string
ABCDEFG
then I want the row with ABC to be displayed.
If my search string is XYZDSDS then the row with value XYZ should be displayed
The answer would be "use LIKE".
See the documentation: https://dev.mysql.com/doc/refman/5.0/en/string-comparison-functions.html
You can do WHERE 'string' LIKE CONCAT(column , '%')
Thus the query becomes:
select * from t1 where 'ABCDEFG' LIKE CONCAT(column1,'%');
If you need to match anywhere in the string:
select * from t1 where 'ABCDEFG' LIKE CONCAT('%',column1,'%');
Here you can see it working in a fiddle:
http://sqlfiddle.com/#!9/d1596/4
Select * from table where #param like '%' + col + '%'
First, you appear to be storing lists of things in a column. This is the wrong approach to storing values in the database. You should have a junction table, with one row per entity and value -- that is, a separate row for ABC and XYZ in your example. SQL has a great data structure for storing lists. It is called a "table", not a "string".
If you are stuck with such a format and using MySQL, there is a function that can help:
where find_in_set('ABC', col)
MySQL treats a comma delimited string as a "set" and offers this function. However, this function cannot use indexes, so it is not particularly efficient. Did I mention that you should use a junction table instead?

nesting sql statements and selecting (MySQL)

Given the following MySQL-statement:
SELECT table_name,column_name
FROM information_schema.columns
WHERE table_schema = 'myschema'
AND column_name REGEXP 'ID$'
ORDER BY table_name,ordinal_position
I would like to select from the result all columns that contain a certain ID #.
Since I don't know the exact name of the column or table I should apply my 2nd SELECT, I need to apply my SELECT to some kind of "placeholders", I think.
E.g., if the resultset of the first request is:
('wp_links', 'link_id')
('wp_options', 'option_id')
('wp_postmeta', 'meta_id')
('wp_postmeta', 'post_id')
then the select should comprise all table_name that are in the first column of the result and should take the column_name of the second column of the result as argument to test whether it contains a certain ID #.
In other words I would like to find all columns in a certain database that are named *_ID and contain a certain ID# and know their corresponding table_name the column belongs to.
You can't do what you're wanting to in SQL but you could do it using procedural SQL, either in a procedure or an anonymous block.
Create a cursor which loops through the results of your statement above. When looping through the results, construct the string you want to execute on each table to see if the desired id exists and execute the query. You will create this string by doing something like:
SET sql_statament = CONCAT("SELECT * FROM ", v_table_name," WHERE ", v_column_name," = ", v_id_number);
This tutorial should help: http://www.mysqltutorial.org/mysql-cursor/
You want to do two different things:
Get information on your database (table and column names).
Get data from your database (values from columns).
You cannot do both at the same time. You can use some programming language firing first an SQL query to get table names and columns and then build a query or several queries to get the data.

Full JOIN MySQL Query is returning empty

So here is a MySQL Query:
SELECT TestSite . * , LoggedCarts . *
FROM TestSite, LoggedCarts
WHERE TestSite.email = 'LoggedCarts.Bill-Email'
LIMIT 0 , 30
It is returning an empty result set, when it should be returning four results based on the tables below.
First Table: LoggedCarts - Column: Bill-Email
casedilla#hotmail.com
crazyandy#theholeintheground.com
Second Table: TestSite - Column: email
samuel#lipsum.com
taco#flavoredkisses.com
honeybadger#dontcare.com
casedilla#hotmail.com
messingwith#sasquatch.com
The goal is to get a MySQL statement that returns the rows in Table: TestSite that don't match the rows in Table: LoggedCarts.
Note: I understand that the use of a hyphen in a column name requires special care when constructing a query, involving backticks to tell MySQL there are special characters. I would change the column names to match up, however the Table: LoggedCarts has data fed via post from a Yahoo Shopping Cart and without heavy preparation before insertion setting the name to anything but the key sent in the post data is daunting.
However, if it turns out rebuilding the data prior to insertion is easier than using a JOIN statement or for some reason using two columns with different names as the comparison columns just doesn't work, I will go through and rebuild the database and PHP code.
Single quotes indicate a string literal. You need to use backticks for identifiers. Also, each component of an identifier must be quoted individually.
SELECT TestSite . * , LoggedCarts . *
FROM TestSite, LoggedCarts
WHERE TestSite.email = LoggedCarts.`Bill-Email`
LIMIT 0 , 30
From the manual:
If any components of a multiple-part name require quoting, quote them individually rather than quoting the name as a whole. For example, write `my-table`.`my-column`, not `my-table.my-column`.
With a bit of research inspired by somne of the hints given, I found the solution I was looking for here: SELECT * WHERE NOT EXISTS
Does exactly what I need it to do, and as a bonus, I like the shorthand syntax that is used that allows you to put in an alias for the table name and use the alias throughout the statement.
SELECT *
FROM TestSite e
WHERE NOT EXISTS
(
SELECT null
FROM LoggedCarts d
WHERE d.`Bill-Email` = e.email
)