According to mysql website i should be able to start a query with an if statement.
IF search_condition THEN statement_list
[ELSEIF search_condition THEN statement_list] ...
[ELSE statement_list]
END IF
but when i try this query
if (count(1)
FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS
WHERE TABLE_SCHEMA = 'dbname'
AND TABLE_NAME='tblname'
AND CONSTRAINT_NAME = 'con_name')
then
alter table table drop foreign key constraint_name;
end if
i get a syntax error saying i have the wrong syntax near "IF" and mysql workbench highlights the if saying syntax error, unexpected if.
i've tried with begin if, and omitting both begin and end if but the error is always the same.
You can't use if or while conditions out side a statement unless they are enclosed in a code block of begin - end. Hence db engine threw an error on your statement.
For your statement to work, you need a stored procedure and some changes to the statement as well.
Example:
delimiter //
drop procedure if exists drop_constraint //
create procedure drop_constraint(
in dbName varchar(64),
in tableName varchar(64),
in constraintName varchar(64) )
begin
declare cnt int default 0;
select count(1) into cnt
from INFORMATION_SCHEMA.TABLE_CONSTRAINTS
where
table_schema = dbName
and table_name = tableName
and constraint_name = constraintName;
-- now check if any found
if ( cnt > 0 ) then -- if found some
-- now, execute your alter statement
-- include your alter table statement here
end if;
end;
//
delimiter ;
Using the above procedure you can check and drop a constraint.
mysql> call drop_constraint( 'test', 'my_table', 'fk_name' );
You cannot, if both table have same structure (Or you put same structure instead of *) you can do this using unions this way
select * from sometable WHERE (SELECT 1 FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_SCHEMA = 'dbname' AND TABLE_NAME='tblname' AND CONSTRAINT_NAME = 'con_name') = 1
union all
select * from anothertable WHERE (SELECT 1 FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE TABLE_SCHEMA = 'dbname' AND TABLE_NAME='tblname' AND CONSTRAINT_NAME = 'con_name') IS NULL
Alternatievly you can achieve this by using stored procedure (With almost same synbtax as you had wrote)
Related
i have error with my if.
ERROR is "IF" is not valid in this position, expecting EOF, ALTER, ANALYZE...
my code looks like this can you help me please :)
IF NOT EXISTS (
SELECT
1
FROM
INFORMATION_SCHEMA.COLUMNS
WHERE
TABLE_NAME = 'clients' AND COLUMN_NAME = 'dateOfRegister')
BEGIN
ALTER TABLE realestate
ADD dateOfRegister DATE NOT NULL
END;
add delimter like this to your request.
delimiter |
IF NOT EXISTS (
SELECT
1
FROM
INFORMATION_SCHEMA.COLUMNS
WHERE
TABLE_NAME = 'clients' AND COLUMN_NAME = 'dateOfRegister')
BEGIN
ALTER TABLE realestate ADD dateOfRegister DATE NOT NULL;
END;
delimiter ;
but it seems that your request will not do what you want to do.
it is better to create a procedure or a function.
I was wondering if it's possible to conditionally run certain statements in MySQL. Something like this:
IF EXISTS (SELECT * FROM information_schema.columns WHERE TABLE_NAME = 'test_table' AND COLUMN_NAME = 'userid' AND IS_NULLABLE = 'NO')
THEN
ALTER TABLE test_table MODIFY userid INT(11) NULL;
END IF;
I've done some googleing and I'm not pulling up anything useful or sane. Wondering if I'm missing something or if this is just a serious MySQL limitation.
Thanks to William_Wilson, Here's my working query. I still think this is totally backwards, but I supposed that's how MySQL rolls...
delimiter //
create procedure update_stuff()
begin
IF EXISTS (SELECT * FROM information_schema.columns WHERE TABLE_NAME = 'test_table' AND COLUMN_NAME = 'userid' AND IS_NULLABLE = 'NO')
THEN
ALTER TABLE test_table MODIFY userid INT(11) NULL;
END IF;
END
//
delimiter ;
-- Execute the procedure
call update_stuff();
-- Drop the procedure
drop procedure update_stuff;
IF NOT EXISTS (SELECT * FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = 'email_subscription' AND COLUMN_NAME = 'subscribe_all')
THEN
ALTER TABLE email_subscription
ADD COLUMN subscribe_all TINYINT(1) DEFAULT 1,
ADD COLUMN subscribe_category varchar(512) DEFAULT NULL;
I had a look at huge amount of examples. but this query doesn't work, I got error of:
ERROR 1064 (42000): You have an error in your SQL syntax; check the
manual that corresponds to your MySQL server version for the right
syntax to use near 'IF NOT EXISTS (SELECT * FROM
INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME =' at line 1
If your host doesn't give you permission to create or run procedures, I think I found another way to do this using PREPARE/EXECUTE and querying the schema:
SET #s = (SELECT IF(
(SELECT COUNT(*)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE table_name = 'table_name'
AND table_schema = DATABASE()
AND column_name = 'col_name'
) > 0,
"SELECT 1",
"ALTER TABLE table_name ADD col_name VARCHAR(100)"
));
PREPARE stmt FROM #s;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
you can create a procedure for the query,
DELIMITER $$
CREATE PROCEDURE Alter_Table()
BEGIN
DECLARE _count INT;
SET _count = ( SELECT COUNT(*)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = 'email_subscription' AND
COLUMN_NAME = 'subscribe_all');
IF _count = 0 THEN
ALTER TABLE email_subscription
ADD COLUMN subscribe_all TINYINT(1) DEFAULT 1,
ADD COLUMN subscribe_category varchar(512) DEFAULT NULL;
END IF;
END $$
DELIMITER ;
There is no equivalent syntax to achieve this in a single MySQL statement.
To get something simlilar, you can either
1) attempt to add the column with an ALTER TABLE, and let MySQL raise an error if a column of that name already exists in the table, or
2) query the information_schema.columns view to check if a column of that name exists in the table.
Note that you really do need to check for the table_schema, as well as the table_name:
SELECT column_name
FROM information_schema.columns
WHERE table_schema = 'foo'
AND table_name = 'email_subscription'
AND column_name = 'subscribe_all'
and based on that, decide whether to run the ALTER TABLE
You are using MS SQL Server syntax in MySQL.
Also add condition for database name to check column existance.
Try this:
DELIMITER $$
CREATE PROCEDURE sp_AlterTable()
BEGIN
IF NOT EXISTS (SELECT * FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_SCHEMA = 'dbName' AND
TABLE_NAME = 'email_subscription' AND
COLUMN_NAME = 'subscribe_all') THEN
ALTER TABLE email_subscription
ADD COLUMN subscribe_all TINYINT(1) DEFAULT 1,
ADD COLUMN subscribe_category VARCHAR(512) DEFAULT NULL;
END IF;
END $$
DELIMITER ;
SET #s = (SELECT IF(
(SELECT COUNT(column_name)
FROM INFORMATION_SCHEMA.COLUMNS
WHERE table_name = 'oc_bitcoin_wallets_receive'
AND table_schema = 'k_opencart2'
AND column_name = 'test3'
) > 0,
"SELECT 1",
"ALTER TABLE `oc_bitcoin_wallets_receive` ADD COLUMN `test3` INT NOT NULL AFTER `test2`;"
));
PREPARE stmt FROM #s;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
Values for edit:
oc_bitcoin_wallets_receive - table name,
k_opencart2 - database name,
test3 - name of new column,
oc_bitcoin_wallets_receive - second location table
test3 - second location column,
test2 - name of column before new column.
Is there a way to check if a column exists in a mySQL DB prior to (or as) the ALTER TABLE ADD coumn_name statement runs? Sort of an IF column DOES NOT EXIST ALTER TABLE thing.
I've tried ALTER IGNORE TABLE my_table ADD my_column but this still throws the error if the column I'm adding already exists.
EDIT: use case is to upgrade a table in an already installed web app-- so to keep things simple, I want to make sure the columns I need exist, and if they don't, add them using ALTER TABLE
Since mysql control statements (e.g. "IF") only work in stored procedures, a temporary one can be created and executed:
DROP PROCEDURE IF EXISTS add_version_to_actor;
DELIMITER $$
CREATE DEFINER=CURRENT_USER PROCEDURE add_version_to_actor ( )
BEGIN
DECLARE colName TEXT;
SELECT column_name INTO colName
FROM information_schema.columns
WHERE table_schema = 'connjur'
AND table_name = 'actor'
AND column_name = 'version';
IF colName is null THEN
ALTER TABLE actor ADD version TINYINT NOT NULL DEFAULT '1' COMMENT 'code version of actor when stored';
END IF;
END$$
DELIMITER ;
CALL add_version_to_actor;
DROP PROCEDURE add_version_to_actor;
Do you think you can try this?:
SELECT IFNULL(column_name, '') INTO #colName
FROM information_schema.columns
WHERE table_name = 'my_table'
AND column_name = 'my_column';
IF #colName = '' THEN
-- ALTER COMMAND GOES HERE --
END IF;
It's no one-liner, but can you at least see if it will work for you? At least while waiting for a better solution..
Utility functions and procedures
First, I have a set of utility functions and procedures that I use to do things like drop foreign keys, normal keys and columns. I just leave them in the database so I can use them as needed.
Here they are.
delimiter $$
create function column_exists(ptable text, pcolumn text)
returns bool
reads sql data
begin
declare result bool;
select
count(*)
into
result
from
information_schema.columns
where
`table_schema` = 'my_database' and
`table_name` = ptable and
`column_name` = pcolumn;
return result;
end $$
create function constraint_exists(ptable text, pconstraint text)
returns bool
reads sql data
begin
declare result bool;
select
count(*)
into
result
from
information_schema.table_constraints
where
`constraint_schema` = 'my_database' and
`table_schema` = 'my_database' and
`table_name` = ptable and
`constraint_name` = pconstraint;
return result;
end $$
create procedure drop_fk_if_exists(ptable text, pconstraint text)
begin
if constraint_exists(ptable, pconstraint) then
set #stat = concat('alter table ', ptable, ' drop foreign key ', pconstraint);
prepare pstat from #stat;
execute pstat;
end if;
end $$
create procedure drop_key_if_exists(ptable text, pconstraint text)
begin
if constraint_exists(ptable, pconstraint) then
set #stat = concat('alter table ', ptable, ' drop key ', pconstraint);
prepare pstat from #stat;
execute pstat;
end if;
end $$
create procedure drop_column_if_exists(ptable text, pcolumn text)
begin
if column_exists(ptable, pcolumn) then
set #stat = concat('alter table ', ptable, ' drop column ', pcolumn);
prepare pstat from #stat;
execute pstat;
end if;
end $$
delimiter ;
Dropping constraints and columns using the utilities above
With those in place, it is pretty easy to use them to check columns and constraints for existence:
-- Drop service.component_id
call drop_fk_if_exists('service', 'fk_service_1');
call drop_key_if_exists('service', 'component_id');
call drop_column_if_exists('service', 'component_id');
-- Drop commit.component_id
call drop_fk_if_exists('commit', 'commit_ibfk_1');
call drop_key_if_exists('commit', 'commit_idx1');
call drop_column_if_exists('commit', 'component_id');
-- Drop component.application_id
call drop_fk_if_exists('component', 'fk_component_1');
call drop_key_if_exists('component', 'application_id');
call drop_column_if_exists('component', 'application_id');
Make a count sentence with the example below by John Watson.
SELECT count(*) FROM information_schema.COLUMNS
WHERE COLUMN_NAME = '...'
and TABLE_NAME = '...'
and TABLE_SCHEMA = '...'
Save that result in an integer and then make it a condition to apply the ADD COLUMN sentence.
You can test if a column exists with:
IF EXISTS (
SELECT * FROM information_schema.COLUMNS
WHERE COLUMN_NAME = '...'
and TABLE_NAME = '...'
and TABLE_SCHEMA = '...')
...
Just fill in your column name, table name, and database name.
Although its quite an old post but still i feel good about sharing my solution to this issue. If column doesn't exist then an exception would occur definitely and then i am creating the column in table.
I just used the code below:
try
{
DATABASE_QUERY="SELECT gender from USER;";
db.rawQuery(DATABASE_QUERY, null);
}
catch (Exception e)
{
e.printStackTrace();
DATABASE_UPGRADE="alter table USER ADD COLUMN gender VARCHAR(10) DEFAULT 0;";
db.execSQL(DATABASE_UPGRADE);
}
You can create a procedure with a CONTINUE handler in case the column exists (please note this code doesn't work in PHPMyAdmin):
DROP PROCEDURE IF EXISTS foo;
CREATE PROCEDURE foo() BEGIN
DECLARE CONTINUE HANDLER FOR 1060 BEGIN END;
ALTER TABLE `tableName` ADD `columnName` int(10) NULL AFTER `otherColumn`;
END;
CALL foo();
DROP PROCEDURE foo;
This code should not raise any error in case the column already exists. It will just do nothing and carry on executing the rest of the SQL.
DELIMITER $$
DROP PROCEDURE IF EXISTS `addcol` $$
CREATE DEFINER=`admin`#`localhost` PROCEDURE `addcol`(tbn varchar(45), cn varchar(45), ct varchar(45))
BEGIN
#tbn: table name, cn: column name, ct: column type
DECLARE CONTINUE HANDLER FOR 1060 BEGIN END;
set cn = REPLACE(cn, ' ','_');
set #a = '';
set #a = CONCAT("ALTER TABLE `", tbn ,"` ADD column `", cn ,"` ", ct);
PREPARE stmt FROM #a;
EXECUTE stmt;
END $$
DELIMITER ;
This syntax work for me :
SHOW COLUMNS FROM < tablename > LIKE '< columnName >'
More in this post :
https://mzulkamal.com/blog/mysql-5-7-check-if-column-exist?viewmode=0
As per MYSQL Community:
IGNORE is a MySQL extension to standard SQL. It controls how ALTER TABLE works if there are duplicates on unique keys in the new table or if warnings occur when strict mode is enabled. If IGNORE is not specified, the copy is aborted and rolled back if duplicate-key errors occur. If IGNORE is specified, only one row is used of rows with duplicates on a unique key. The other conflicting rows are deleted. Incorrect values are truncated to the closest matching acceptable value.
So a working Code is:
ALTER IGNORE TABLE CLIENTS ADD CLIENT_NOTES TEXT DEFAULT NULL;
Data posted here:
http://dev.mysql.com/doc/refman/5.1/en/alter-table.html
Our system's automated database migration process involves running .sql scripts containing new table definitions and their accompanying indexes.
I require the ability to create these tables and indexes only if they don't already exist. Tables are taken care of by using IF NOT EXISTS but no such syntax exists when creating indexes.
I've tried to write a stored procedure, shown below, but this fails presumably as you can't select from a show statement.
DELIMITER $$
DROP PROCEDURE IF EXISTS csi_add_index $$
CREATE PROCEDURE csi_add_index(in theTable varchar(128), in theIndexName varchar(128), in theIndexColumns varchar(128) )
BEGIN
IF(((SELECT COUNT(*) FROM (SHOW KEYS FROM theTable WHERE key_name = theIndexName)) tableInfo = 0) THEN
SET #s = CONCAT('CREATE INDEX ' , theIndexName , ' ON ' , theTable, '(', theIndexColumns, ')');
PREPARE stmt FROM #s;
EXECUTE stmt;
END IF;
END $$
I've considered dropping and recreating but the process, as it exists, assumes that it'll encounter no errors hence me wanting to check for existence first.
Is there another way to retrieve the indexes of a table to check if an index already exists before creating or can anyone suggest a better approach to managing this?
EDIT: Please note that this is an automated procedure, no human intervention.
SELECT INDEX_NAME FROM INFORMATION_SCHEMA.STATISTICS WHERE
`TABLE_CATALOG` = 'def' AND `TABLE_SCHEMA` = DATABASE() AND
`TABLE_NAME` = theTable AND `INDEX_NAME` = theIndexName
After some more banging my head off the wall and intense googling I found the information_schema.statistics table. This contains the index_name for a table.
My stored procedure is now
DELIMITER $$
DROP PROCEDURE IF EXISTS csi_add_index $$
CREATE PROCEDURE csi_add_index(in theTable varchar(128), in theIndexName varchar(128), in theIndexColumns varchar(128) )
BEGIN
IF((SELECT COUNT(*) AS index_exists FROM information_schema.statistics WHERE TABLE_SCHEMA = DATABASE() and table_name =
theTable AND index_name = theIndexName) = 0) THEN
SET #s = CONCAT('CREATE INDEX ' , theIndexName , ' ON ' , theTable, '(', theIndexColumns, ')');
PREPARE stmt FROM #s;
EXECUTE stmt;
END IF;
END $$
and works as expected.
Thanks for the suggestions.
Use SHOW INDEX FROM mytable FROM mydb; and check if the index is present - each of the returned rows represents one part of an index; the column that would probably interest you most is Key_name, as it contains the name of the index. Documentation here.
You can query infomration_schema database for this and many more useful information
http://dev.mysql.com/doc/refman/5.0/en/information-schema.html
Since Text and Blobs need a size, I added it to the stored procedure.
DROP PROCEDURE IF EXISTS createIndex;
DELIMITER $$
-- Create a temporary stored procedure for checking if Indexes exist before attempting to re-create them. => can be called
$$
CREATE PROCEDURE `createIndex`(
IN `tableName` VARCHAR(128),
IN `indexName` VARCHAR(128),
IN `indexColumns` VARCHAR(128),
IN `indexSize` VARCHAR(128)
)
BEGIN
IF((SELECT COUNT(*) AS index_exists FROM information_schema.statistics WHERE TABLE_SCHEMA = DATABASE() AND table_name = tableName AND index_name = indexName) = 0) THEN
IF(indexSize > 0) THEN
SET #sqlCommand = CONCAT('CREATE INDEX ' , indexName , ' ON ' , tableName, '(', indexColumns, '(' , indexSize, '))');
ELSE
SET #sqlCommand = CONCAT('CREATE INDEX ' , indexName , ' ON ' , tableName, '(', indexColumns, ')');
END IF;
PREPARE _preparedStatement FROM #sqlCommand;
EXECUTE _preparedStatement;
END IF;
END $$
DELIMITER ;
You could just create another table with the correct indices, copy everything from the old table and then drop it and rename the new table back to what the old one used to be. A bit hackish and might be a bit heavy for big tables but still fairly straightforward.
Not a new version but a more complete solution that includes a call to create 2 indexes.
USE MyDatabaseName;
DELIMITER $$
-- Create a temporary stored procedure for checking if Indexes exist before attempting to re-create them.
DROP PROCEDURE IF EXISTS `MyDatabaseName`.`spCreateIndex` $$
CREATE PROCEDURE `MyDatabaseName`.`spCreateIndex` (tableName VARCHAR(128), in indexName VARCHAR(128), in indexColumns VARCHAR(128))
BEGIN
IF((SELECT COUNT(*) AS index_exists FROM information_schema.statistics WHERE TABLE_SCHEMA = DATABASE() AND table_name = tableName AND index_name = indexName) = 0) THEN
SET #sqlCommand = CONCAT('CREATE INDEX ' , indexName , ' ON ' , tableName, '(', indexColumns, ')');
PREPARE _preparedStatement FROM #sqlCommand;
EXECUTE _preparedStatement;
END IF;
END $$
DELIMITER ;
-- Create the Indexes if they do not exist already.
CALL spCreateIndex('MyCustomers', 'idxCustNum', 'CustomerNumber');
CALL spCreateIndex('MyProducts', 'idxProductName', 'ProductName');
DELIMITER $$
-- Drop the temporary stored procedure.
DROP PROCEDURE IF EXISTS `MyDatabaseName`.`spCreateIndex` $$
DELIMITER ;