Related
I'm trying to figure out how to create a single MySQL query that will allow me to display only the next 10 characters following the string "filter" in the Message field. The string "filter" appears at various positions in each record, so I can't use a position filter.
I've been trying to use something like like what I have below, however I've been unable to get the correct query.
SELECT RIGHT(Message,LOCATE('filter',Message) - 10) FROM table
The Message field records within the table looks like:
QgySSW8fwD25iQ.filter0019p3las1-31205-59C3D
6t2fJw.filter0010p3las1-9745-59
filter0025p3las1-13130-59C3D317
And I'm looking for them to look like this after the query:
0019p3las1
0010p3las1
0025p3las1
Any help is greatly appreciated.
Use a combination of LOCATE() within SUBSTRING(). See this SQL Fiddle
CREATE TABLE Table1
(`message` varchar(200))
;
INSERT INTO Table1
(`message`)
VALUES
('QgySSW8fwD25iQ.filter0019p3las1-31205-59C3D'),
('6t2fJw.filter0010p3las1-9745-59'),
('filter0025p3las1-13130-59C3D317')
;
Query 1:
select
SUBSTRING(message,LOCATE('filter',Message)+6,10)
from table1
Note that the +6 is to offset for the length of "filter" because LOCATE finds the position of the "f" and you then need to add 6 for the other characters "ilter". Once that number is determined then just get the next 10 characters.
Results:
| SUBSTRING(message,LOCATE('filter',Message)+6,10) |
|--------------------------------------------------|
| 0019p3las1 |
| 0010p3las1 |
| 0025p3las1 |
See SQLFiddle.
Result table structure
Create table resulttbl (
id int(6) primary key auto_increment ,
resultFIlter varchar(1000)
);
Function to split string
CREATE FUNCTION strSplit(x VARCHAR(65000), delim VARCHAR(12), pos INTEGER)
RETURNS VARCHAR(65000)
BEGIN
DECLARE output VARCHAR(65000);
SET output = REPLACE(SUBSTRING(SUBSTRING_INDEX(x, delim, pos)
, LENGTH(SUBSTRING_INDEX(x, delim, pos - 1)) + 1)
, delim
, '');
IF output = '' THEN SET output = null; END IF;
RETURN output;
END;
Stored procedure to split and insert into result table
CREATE PROCEDURE FilterTable()
BEGIN
DECLARE i INTEGER;
DECLARE endpos INTEGER;
DECLARE fullstr VARCHAR(1000);
DECLARE result VARCHAR(1000);
SET fullstr = 'QgySSW8fwD25iQ.filter0019p3las1-31205-59C3D 6t2fJw.filter0010p3las1-9745-59 filter0025p3las1-13130-59C3D317';
SET i = 2;
SET endpos=LENGTH(fullstr) - LENGTH(REPLACE(fullstr, 'filter', '')) ;
delete from resulttbl;
REPEAT
SET result=strSplit(fullstr, 'filter', i);
IF result IS NOT NULL THEN
SET result=LEFT(result,10);
INSERT INTO resulttbl (resultFIlter) values(result);
END IF;
SET i = i + 1;
UNTIL i >= endpos
END REPEAT;
END ;
Call the procedure using the statement CALL FilterTable().
Now the result of your procedure is available on the table resulttbl.
You can get the values from that table using select statement as SELECT * from resulttbl.
Result
id resultFIlter
1 0019p3las1
2 0010p3las1
3 0025p3las1
I'm looking to find records in a table that match a specific number that the user enters. So, the user may enter 12345, but this could be 123zz4-5 in the database.
I imagine something like this would work, if PHP functions worked in MySQL.
SELECT * FROM foo WHERE preg_replace("/[^0-9]/","",bar) = '12345'
What's the equivalent function or way to do this with just MySQL?
Speed is not important.
I realise that this is an ancient topic but upon googling this problem I couldn't find a simple solution (I saw the venerable agents but think this is a simpler solution) so here's a function I wrote, seems to work quite well.
DROP FUNCTION IF EXISTS STRIP_NON_DIGIT;
DELIMITER $$
CREATE FUNCTION STRIP_NON_DIGIT(input VARCHAR(255))
RETURNS VARCHAR(255)
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
WHILE iterator < (LENGTH(input) + 1) DO
IF SUBSTRING(input, iterator, 1) IN ( '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ) THEN
SET output = CONCAT(output, SUBSTRING(input, iterator, 1));
END IF;
SET iterator = iterator + 1;
END WHILE;
RETURN output;
END
$$
You can easily do what you want with REGEXP_REPLACE (compatible with MySQL 8+ and MariaDB 10.0.5+)
REGEXP_REPLACE(expr, pat, repl[, pos[, occurrence[, match_type]]])
Replaces occurrences in the string expr that match the regular expression specified by the pattern pat with the replacement string repl, and returns the resulting string. If expr, pat, or repl is NULL, the return value is NULL.
Go to REGEXP_REPLACE doc: MySQL or MariaDB
Try it:
SELECT REGEXP_REPLACE('123asd12333', '[a-zA-Z]+', '');
Output:
12312333
Updated 2022: as per Marlom's answer, you can now use REGEX_REPLACE - which will perform even better than my historical answer here.
Most upvoted answer above isn't the fastest.
Full kudos to them for giving a working proposal to bounce off!
This is an improved version:
DELIMITER ;;
DROP FUNCTION IF EXISTS `STRIP_NON_DIGIT`;;
CREATE DEFINER=`root`#`localhost` FUNCTION `STRIP_NON_DIGIT`(input VARCHAR(255)) RETURNS VARCHAR(255) CHARSET utf8
READS SQL DATA
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
DECLARE lastDigit INT DEFAULT 1;
DECLARE len INT;
SET len = LENGTH(input) + 1;
WHILE iterator < len DO
-- skip past all digits
SET lastDigit = iterator;
WHILE ORD(SUBSTRING(input, iterator, 1)) BETWEEN 48 AND 57 AND iterator < len DO
SET iterator = iterator + 1;
END WHILE;
IF iterator != lastDigit THEN
SET output = CONCAT(output, SUBSTRING(input, lastDigit, iterator - lastDigit));
END IF;
WHILE ORD(SUBSTRING(input, iterator, 1)) NOT BETWEEN 48 AND 57 AND iterator < len DO
SET iterator = iterator + 1;
END WHILE;
END WHILE;
RETURN output;
END;;
Testing 5000 times on a test server:
-- original
Execution Time : 7.389 sec
Execution Time : 7.257 sec
Execution Time : 7.506 sec
-- ORD between not string IN
Execution Time : 4.031 sec
-- With less substrings
Execution Time : 3.243 sec
Execution Time : 3.415 sec
Execution Time : 2.848 sec
While it's not pretty and it shows results that don't match, this helps:
SELECT * FROM foo WHERE bar LIKE = '%1%2%3%4%5%'
I would still like to find a better solution similar to the item in the original question.
There's no regexp replace, only a plain string REPLACE().
MySQL has the REGEXP operator, but it's only a match tester not a replacer, so you would have to turn the logic inside-out:
SELECT * FROM foo WHERE bar REGEXP '[^0-9]*1[^0-9]*2[^0-9]*3[^0-9]*4[^0-9]*5[^0-9]*';
This is like your version with LIKE but matches more accurately. Both will perform equally badly, needing a full table scan without indexes.
On MySQL 8.0+ there is a new native function called REGEXP_REPLACE. A clean solution to this question would be:
SELECT * FROM foo WHERE REGEXP_REPLACE(bar,'[^0-9]+',"") = '12345'
The simplest way I can think to do it is to use the MySQL REGEXP operator a la:
WHERE foo LIKE '1\D*2\D*3\D*4\D*5'
It's not especially pretty but MySQL doesn't have a preg_replace function so I think it's the best you're going to get.
Personally, if this only-numeric data is so important, I'd keep a separate field just to contain the stripped data. It'll make your lookups a lot faster than with the regular expression search.
This blog post details how to strip non-numeric characters from a string via a MySQL function:
SELECT NumericOnly("asdf11asf");
returns 11
http://venerableagents.wordpress.com/2011/01/29/mysql-numeric-functions/
There's no regex replace as far as I'm concerned, but I found this solution;
--Create a table with numbers
DROP TABLE IF EXISTS ints;
CREATE TABLE ints (i INT UNSIGNED NOT NULL PRIMARY KEY);
INSERT INTO ints (i) VALUES
( 1), ( 2), ( 3), ( 4), ( 5), ( 6), ( 7), ( 8), ( 9), (10),
(11), (12), (13), (14), (15), (16), (17), (18), (19), (20);
--Then extract the numbers from the specified column
SELECT
bar,
GROUP_CONCAT(SUBSTRING(bar, i, 1) ORDER BY i SEPARATOR '')
FROM foo
JOIN ints ON i BETWEEN 1 AND LENGTH(bar)
WHERE
SUBSTRING(bar, i, 1) IN ('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')
GROUP BY bar;
It works for me and I use MySQL 5.0
Also I found this place that could help.
I have a similar situation, matching products to barcodes where the barcode doesn't store none alpha numerics sometimes, so 102.2234 in the DB needs to be found when searching for 1022234.
In the end I just added a new field, reference_number to the products tables, and have php strip out the none alpha numerics in the product_number to populate reference_number whenever a new products is added.
You'd need to do a one time scan of the table to create all the reference_number fields for existing products.
You can then setup your index, even if speed is not a factor for this operation, it is still a good idea to keep the database running well so this query doesn't bog it down and slow down other queries.
I came across this solution. The top answer by user1467716 will work in phpMyAdmin with a small change: add a second delimiter tag to the end of the code.
phpMyAdmin version is 4.1.14; MySQL version 5.6.20
I also added a length limiter using
DECLARE count INT DEFAULT 0; in the declarations
AND count < 5 in the WHILE statement
SET COUNT=COUNT+1; in the IF statement
Final form:
DROP FUNCTION IF EXISTS STRIP_NON_DIGIT;
DELIMITER $$
CREATE FUNCTION STRIP_NON_DIGIT(input VARCHAR(255))
RETURNS VARCHAR(255)
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
DECLARE count INT DEFAULT 0;
WHILE iterator < (LENGTH(input) + 1) AND count < 5 DO --limits to 5 chars
IF SUBSTRING(input, iterator, 1) IN ( '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ) THEN
SET output = CONCAT(output, SUBSTRING(input, iterator, 1));
SET COUNT=COUNT+1;
END IF;
SET iterator = iterator + 1;
END WHILE;
RETURN output;
END
$$
DELIMITER $$ --added this
How big is table with foo? If it is small, and speed really doesn't matter, you might pull the row ID and foo, loop over it using the PHP replace functions to compare, and then pull the info you want by row number.
Of course, if the table is too big, this won't work well.
try this example. this is used for phone numbers, however you can modify it for your needs.
-- function removes non numberic characters from input
-- returne only the numbers in the string
CREATE DEFINER =`root`#`localhost` FUNCTION `remove_alpha`(inputPhoneNumber VARCHAR(50))
RETURNS VARCHAR(50)
CHARSET latin1
DETERMINISTIC
BEGIN
DECLARE inputLenght INT DEFAULT 0;
-- var for our iteration
DECLARE counter INT DEFAULT 1;
-- if null is passed, we still return an tempty string
DECLARE sanitizedText VARCHAR(50) DEFAULT '';
-- holder of each character during the iteration
DECLARE oneChar VARCHAR(1) DEFAULT '';
-- we'll process only if it is not null.
IF NOT ISNULL(inputPhoneNumber)
THEN
SET inputLenght = LENGTH(inputPhoneNumber);
WHILE counter <= inputLenght DO
SET oneChar = SUBSTRING(inputPhoneNumber, counter, 1);
IF (oneChar REGEXP ('^[0-9]+$'))
THEN
SET sanitizedText = Concat(sanitizedText, oneChar);
END IF;
SET counter = counter + 1;
END WHILE;
END IF;
RETURN sanitizedText;
END
to use this user defined function (UDF).
let's say you have a column of phone numbers:
col1
(513)983-3983
1-838-338-9898
phone983-889-8383
select remove_alpha(col1) from mytable
The result would be;
5139833983
18383389898
9838898383
thought I would share this since I built it off the function from here. I rearranged just so I can read it easier (I'm just server side).
You call it by passing in a table name and column name to have it strip all existing non-numeric characters from that column. I inherited a lot of bad table structures that put a ton of int fields as varchar so I needed a way to clean these up quickly before I can modify the column to an integer.
drop procedure if exists strip_non_numeric_characters;
DELIMITER ;;
CREATE PROCEDURE `strip_non_numeric_characters`(
tablename varchar(100)
,columnname varchar(100)
)
BEGIN
-- =============================================
-- Author: <Author,,David Melton>
-- Create date: <Create Date,,2/26/2019>
-- Description: <Description,,loops through data and strips out the bad characters in whatever table and column you pass it>
-- =============================================
#this idea was generated from the idea STRIP_NON_DIGIT function
#https://stackoverflow.com/questions/287105/mysql-strip-non-numeric-characters-to-compare
declare input,output varchar(255);
declare iterator,lastDigit,len,counter int;
declare date_updated varchar(100);
select column_name
into date_updated
from information_schema.columns
where table_schema = database()
and extra rlike 'on update CURRENT_TIMESTAMP'
and table_name = tablename
limit 1;
#only goes up to 255 so people don't run this for a longtext field
#just to be careful, i've excluded columns that are part of keys, that could potentially mess something else up
set #find_column_length =
concat("select character_maximum_length
into #len
from information_schema.columns
where table_schema = '",database(),"'
and column_name = '",columnname,"'
and table_name = '",tablename,"'
and length(ifnull(character_maximum_length,100)) < 255
and data_type in ('char','varchar')
and column_key = '';");
prepare stmt from #find_column_length;
execute stmt;
deallocate prepare stmt;
set counter = 1;
set len = #len;
while counter <= ifnull(len,1) DO
#this just removes it by putting all the characters before and after the character i'm looking at
#you have to start at the end of the field otherwise the lengths don't stay in order and you have to run it multiple times
set #update_query =
concat("update `",tablename,"`
set `",columnname,"` = concat(substring(`",columnname,"`,1,",len - counter,"),SUBSTRING(`",columnname,"`,",len - counter,",",counter - 1,"))
",if(date_updated is not null,concat(",`",date_updated,"` = `",date_updated,"`
"),''),
"where SUBSTRING(`",columnname,"`,",len - counter,", 1) not REGEXP '^[0-9]+$';");
prepare stmt from #update_query;
execute stmt;
deallocate prepare stmt;
set counter = counter + 1;
end while;
END ;;
DELIMITER ;
To search for numbers that match a particular numeric pattern in a string, first remove all the alphabets and special characters in a similar manner as below then convert the value to an integer and then search
SELECT *
FROM foo
WHERE Convert(Regexp_replace(bar, '[a-zA-Z]+', ''), signed) = 12345
I think you don't need complicated functions for that.
I've found a REGEXP_REPLACE solution using built-in mysql character class names. You can read about them in a table in the docs. Basically, they are mysql-specific names for commonly matched groups of characters like [:alnum:] for alpha-numeric characters, [:alpha:] for only alphabetic characters and so on.
So my version of the REGEXP_REPLACE:
REGEXP_REPLACE('My number is: +59 (29) 889-23-56', '[[:alpha:][:blank:][:punct:][:cntrl:]]', '')
will yield 59298892356 as per requirements.
Being someone that has version 5.7, doesn't have the privilege to create functions, and it's not practical to bring my data down into my code, I found Nelson Miranda's Answer amazing. I wanted to share it in a subquery version which I found more useful.
DROP TABLE IF EXISTS ints;
CREATE TABLE ints (i INT UNSIGNED NOT NULL PRIMARY KEY);
INSERT INTO ints (i) VALUES
( 1), ( 2), ( 3), ( 4), ( 5), ( 6), ( 7), ( 8), ( 9), (10),
(11), (12), (13), (14), (15), (16), (17), (18), (19), (20);
SELECT * FROM foo f
WHERE (SELECT GROUP_CONCAT(SUBSTRING(f.bar, i, 1) ORDER BY i SEPARATOR '')
FROM ints
WHERE i BETWEEN 1 AND LENGTH(f.bar)
AND SUBSTRING(f.bar, i, 1) IN ('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')) = '12345'
Note: The ints table can be a temporary table.
If you are on MySQL 5.7 or below, and you just need something quick and dirty without having to define a new function, and you have a small number of known non-numerics to filter out, something like this can work well...
Select replace(replace(replace(replace(replace(phone, '+1', ''), '(', ''), ')', ''), '-', ''), ' ', '') from customers;
i would like to replace value like:
1106,1107,1108
from select query to a string link like:
http://something.com/img/1/1/0/6/1106.jpg,http://something.com/img/1/1/0/7/1107.jpg,http://something.com/img/1/1/0/8/1108.jpg
can it be done in mysql query?
Assuming the image names have variable length, I think you'll need to write a stored function to implement this natively in MySQL, there's no obvious built-in function.
The example below will take 1106 and convert it to http://something.com/img/1/1/0/6/1106.jpg. To parse multiple image IDs like 1106,1107,1108 you'd need to extend it to insert the path again every time it finds a comma, or (better) select the results out of the database in a way that is not comma-separated.
DELIMITER //
CREATE FUNCTION TO_IMAGE_PATH(id VARCHAR(255), path VARCHAR(255))
RETURNS VARCHAR(255) DETERMINISTIC NO SQL
BEGIN
DECLARE output VARCHAR(255) DEFAULT path;
DECLARE position INT DEFAULT 1;
WHILE position <= LENGTH(id) DO
SET output = CONCAT(output, SUBSTRING(id, position, 1), '/');
SET position = position + 1;
END WHILE;
SET output = CONCAT(output, id, '.jpg');
RETURN output;
END//
DELIMITER ;
SELECT TO_IMAGE_PATH('1106', 'http://something.com/img/');
-- Output: http://something.com/img/1/1/0/6/1106.jpg
You might prefer to pass in the jpg extension, or hard-code the initial path.
While this does work, this seems like an example of a problem which might be better solved in another programming language after you have selected out your results.
If all of the image IDs are exactly 4 digits long, you could do the simpler (but less elegant)
SELECT CONCAT(
'http://something.com/img/',
SUBSTRING(field_name, 1, 1), '/',
SUBSTRING(field_name, 2, 1), '/',
SUBSTRING(field_name, 3, 1), '/',
SUBSTRING(field_name, 4, 1), '/',
field_name, '.jpg');
Again, you'd need to work out how to select the values out so they aren't comma-separated. In general, if you're storing values comma-separated in your database, then you shouldn't be.
It appears that MySQL doesn't have array variables. What should I use instead?
There seem to be two alternatives suggested: A set-type scalar and temporary tables. The question I linked to suggests the former. But is it good practice to use these instead of array variables? Alternatively, if I go with sets, what would be the set-based idiom equivalent to foreach?
Well, I've been using temporary tables instead of array variables. Not the greatest solution, but it works.
Note that you don't need to formally define their fields, just create them using a SELECT:
DROP TEMPORARY TABLE IF EXISTS my_temp_table;
CREATE TEMPORARY TABLE my_temp_table
SELECT first_name FROM people WHERE last_name = 'Smith';
(See also Create temporary table from select statement without using Create Table.)
You can achieve this in MySQL using WHILE loop:
SET #myArrayOfValue = '2,5,2,23,6,';
WHILE (LOCATE(',', #myArrayOfValue) > 0)
DO
SET #value = ELT(1, #myArrayOfValue);
SET #myArrayOfValue= SUBSTRING(#myArrayOfValue, LOCATE(',',#myArrayOfValue) + 1);
INSERT INTO `EXEMPLE` VALUES(#value, 'hello');
END WHILE;
EDIT:
Alternatively you can do it using UNION ALL:
INSERT INTO `EXEMPLE`
(
`value`, `message`
)
(
SELECT 2 AS `value`, 'hello' AS `message`
UNION ALL
SELECT 5 AS `value`, 'hello' AS `message`
UNION ALL
SELECT 2 AS `value`, 'hello' AS `message`
UNION ALL
...
);
Try using FIND_IN_SET() function of MySql
e.g.
SET #c = 'xxx,yyy,zzz';
SELECT * from countries
WHERE FIND_IN_SET(countryname,#c);
Note: You don't have to SET variable in StoredProcedure if you are passing parameter with CSV values.
Nowadays using a JSON array would be an obvious answer.
Since this is an old but still relevant question I produced a short example.
JSON functions are available since mySQL 5.7.x / MariaDB 10.2.3
I prefer this solution over ELT() because it's really more like an array and this 'array' can be reused in the code.
But be careful: It (JSON) is certainly much slower than using a temporary table. Its just more handy. imo.
Here is how to use a JSON array:
SET #myjson = '["gmail.com","mail.ru","arcor.de","gmx.de","t-online.de",
"web.de","googlemail.com","freenet.de","yahoo.de","gmx.net",
"me.com","bluewin.ch","hotmail.com","hotmail.de","live.de",
"icloud.com","hotmail.co.uk","yahoo.co.jp","yandex.ru"]';
SELECT JSON_LENGTH(#myjson);
-- result: 19
SELECT JSON_VALUE(#myjson, '$[0]');
-- result: gmail.com
And here a little example to show how it works in a function/procedure:
DELIMITER //
CREATE OR REPLACE FUNCTION example() RETURNS varchar(1000) DETERMINISTIC
BEGIN
DECLARE _result varchar(1000) DEFAULT '';
DECLARE _counter INT DEFAULT 0;
DECLARE _value varchar(50);
SET #myjson = '["gmail.com","mail.ru","arcor.de","gmx.de","t-online.de",
"web.de","googlemail.com","freenet.de","yahoo.de","gmx.net",
"me.com","bluewin.ch","hotmail.com","hotmail.de","live.de",
"icloud.com","hotmail.co.uk","yahoo.co.jp","yandex.ru"]';
WHILE _counter < JSON_LENGTH(#myjson) DO
-- do whatever, e.g. add-up strings...
SET _result = CONCAT(_result, _counter, '-', JSON_VALUE(#myjson, CONCAT('$[',_counter,']')), '#');
SET _counter = _counter + 1;
END WHILE;
RETURN _result;
END //
DELIMITER ;
SELECT example();
Dont know about the arrays, but there is a way to store comma-separated lists in normal VARCHAR column.
And when you need to find something in that list you can use the FIND_IN_SET() function.
I know that this is a bit of a late response, but I recently had to solve a similar problem and thought that this may be useful to others.
Background
Consider the table below called 'mytable':
The problem was to keep only latest 3 records and delete any older records whose systemid=1 (there could be many other records in the table with other systemid values)
It would be good if you could do this simply using the statement
DELETE FROM mytable WHERE id IN (SELECT id FROM `mytable` WHERE systemid=1 ORDER BY id DESC LIMIT 3)
However this is not yet supported in MySQL and if you try this then you will get an error like
...doesn't yet support 'LIMIT & IN/ALL/SOME subquery'
So a workaround is needed whereby an array of values is passed to the IN selector using variable. However, as variables need to be single values, I would need to simulate an array. The trick is to create the array as a comma separated list of values (string) and assign this to the variable as follows
SET #myvar = (SELECT GROUP_CONCAT(id SEPARATOR ',') AS myval FROM (SELECT * FROM `mytable` WHERE systemid=1 ORDER BY id DESC LIMIT 3 ) A GROUP BY A.systemid);
The result stored in #myvar is
5,6,7
Next, the FIND_IN_SET selector is used to select from the simulated array
SELECT * FROM mytable WHERE FIND_IN_SET(id,#myvar);
The combined final result is as follows:
SET #myvar = (SELECT GROUP_CONCAT(id SEPARATOR ',') AS myval FROM (SELECT * FROM `mytable` WHERE systemid=1 ORDER BY id DESC LIMIT 3 ) A GROUP BY A.systemid);
DELETE FROM mytable WHERE FIND_IN_SET(id,#myvar);
I am aware that this is a very specific case. However it can be modified to suit just about any other case where a variable needs to store an array of values.
I hope that this helps.
DELIMITER $$
CREATE DEFINER=`mysqldb`#`%` PROCEDURE `abc`()
BEGIN
BEGIN
set #value :='11,2,3,1,';
WHILE (LOCATE(',', #value) > 0) DO
SET #V_DESIGNATION = SUBSTRING(#value,1, LOCATE(',',#value)-1);
SET #value = SUBSTRING(#value, LOCATE(',',#value) + 1);
select #V_DESIGNATION;
END WHILE;
END;
END$$
DELIMITER ;
Maybe create a temporary memory table with columns (key, value) if you want associative arrays. Having a memory table is the closest thing to having arrays in mysql
Here’s how I did it.
First, I created a function that checks whether a Long/Integer/whatever value is in a list of values separated by commas:
CREATE DEFINER = 'root'#'localhost' FUNCTION `is_id_in_ids`(
`strIDs` VARCHAR(255),
`_id` BIGINT
)
RETURNS BIT(1)
NOT DETERMINISTIC
CONTAINS SQL
SQL SECURITY DEFINER
COMMENT ''
BEGIN
DECLARE strLen INT DEFAULT 0;
DECLARE subStrLen INT DEFAULT 0;
DECLARE subs VARCHAR(255);
IF strIDs IS NULL THEN
SET strIDs = '';
END IF;
do_this:
LOOP
SET strLen = LENGTH(strIDs);
SET subs = SUBSTRING_INDEX(strIDs, ',', 1);
if ( CAST(subs AS UNSIGNED) = _id ) THEN
-- founded
return(1);
END IF;
SET subStrLen = LENGTH(SUBSTRING_INDEX(strIDs, ',', 1));
SET strIDs = MID(strIDs, subStrLen+2, strLen);
IF strIDs = NULL or trim(strIds) = '' THEN
LEAVE do_this;
END IF;
END LOOP do_this;
-- not founded
return(0);
END;
So now you can search for an ID in a comma-separated list of IDs, like this:
select `is_id_in_ids`('1001,1002,1003',1002);
And you can use this function inside a WHERE clause, like this:
SELECT * FROM table1 WHERE `is_id_in_ids`('1001,1002,1003',table1_id);
This was the only way I found to pass an "array" parameter to a PROCEDURE.
I'm surprised none of the answers mention ELT/FIELD.
ELT/FIELD works very similar to an array especially if you have static data.
FIND_IN_SET also works similar but doesn't have a built in complementary
function but it's easy enough to write one.
mysql> select elt(2,'AA','BB','CC');
+-----------------------+
| elt(2,'AA','BB','CC') |
+-----------------------+
| BB |
+-----------------------+
1 row in set (0.00 sec)
mysql> select field('BB','AA','BB','CC');
+----------------------------+
| field('BB','AA','BB','CC') |
+----------------------------+
| 2 |
+----------------------------+
1 row in set (0.00 sec)
mysql> select find_in_set('BB','AA,BB,CC');
+------------------------------+
| find_in_set('BB','AA,BB,CC') |
+------------------------------+
| 2 |
+------------------------------+
1 row in set (0.00 sec)
mysql> SELECT SUBSTRING_INDEX(SUBSTRING_INDEX('AA,BB,CC',',',2),',',-1);
+-----------------------------------------------------------+
| SUBSTRING_INDEX(SUBSTRING_INDEX('AA,BB,CC',',',2),',',-1) |
+-----------------------------------------------------------+
| BB |
+-----------------------------------------------------------+
1 row in set (0.01 sec)
Is an array variable really necessary?
I ask because I originally landed here wanting to add an array as a MySQL table variable. I was relatively new to database design and trying to think of how I'd do it in a typical programming language fashion.
But databases are different. I thought I wanted an array as a variable, but it turns out that's just not a common MySQL database practice.
Standard Practice
The alternative solution to arrays is to add an additional table, and then reference your original table with a foreign key.
As an example, let's imagine an application that keeps track of all the items every person in a household wants to buy at the store.
The commands for creating the table I originally envisioned would have looked something like this:
#doesn't work
CREATE TABLE Person(
name VARCHAR(50) PRIMARY KEY
buy_list ARRAY
);
I think I envisioned buy_list to be a comma-separated string of items or something like that.
But MySQL doesn't have an array type field, so I really needed something like this:
CREATE TABLE Person(
name VARCHAR(50) PRIMARY KEY
);
CREATE TABLE BuyList(
person VARCHAR(50),
item VARCHAR(50),
PRIMARY KEY (person, item),
CONSTRAINT fk_person FOREIGN KEY (person) REFERENCES Person(name)
);
Here we define a constraint named fk_person. It says that the 'person' field in BuyList is a foreign key. In other words, it's a primary key in another table, specifically the 'name' field in the Person table, which is what REFERENCES denotes.
We also defined the combination of person and item to be the primary key, but technically that's not necessary.
Finally, if you want to get all the items on a person's list, you can run this query:
SELECT item FROM BuyList WHERE person='John';
This gives you all the items on John's list. No arrays necessary!
This is my solution to use a variable containing a list of elements.
You can use it in simple queries (no need to use store procedures or create tables).
I found somewhere else on the site the trick to use the JSON_TABLE function (it works in mysql 8, I dunno of it works in other versions).
set #x = '1,2,3,4' ;
select c.NAME
from colors c
where
c.COD in (
select *
from json_table(
concat('[',#x,']'),
'$[*]' columns (id int path '$') ) t ) ;
Also, you may need to manage the case of one or more variables set to empty_string.
In this case I added another trick (the query does not return error even if x, y, or both x and y are empty strings):
set #x = '' ;
set #y = 'yellow' ;
select c.NAME
from colors
where
if(#y = '', 1 = 1, c.NAME = #y)
and if(#x = '', 1, c.COD) in (
select *
from json_table(
concat('[',if(#x = '', 1, #x),']'),
'$[*]' columns (id int path '$') ) t) ;
This works fine for list of values:
SET #myArrayOfValue = '2,5,2,23,6,';
WHILE (LOCATE(',', #myArrayOfValue) > 0)
DO
SET #value = ELT(1, #myArrayOfValue);
SET #STR = SUBSTRING(#myArrayOfValue, 1, LOCATE(',',#myArrayOfValue)-1);
SET #myArrayOfValue = SUBSTRING(#myArrayOfValue, LOCATE(',', #myArrayOfValue) + 1);
INSERT INTO `Demo` VALUES(#STR, 'hello');
END WHILE;
Both versions using sets didn't work for me (tested with MySQL 5.5). The function ELT() returns the whole set. Considering the WHILE statement is only avaible in PROCEDURE context i added it to my solution:
DROP PROCEDURE IF EXISTS __main__;
DELIMITER $
CREATE PROCEDURE __main__()
BEGIN
SET #myArrayOfValue = '2,5,2,23,6,';
WHILE (LOCATE(',', #myArrayOfValue) > 0)
DO
SET #value = LEFT(#myArrayOfValue, LOCATE(',',#myArrayOfValue) - 1);
SET #myArrayOfValue = SUBSTRING(#myArrayOfValue, LOCATE(',',#myArrayOfValue) + 1);
END WHILE;
END;
$
DELIMITER ;
CALL __main__;
To be honest, i don't think this is a good practice. Even if its realy necessary, this is barely readable and quite slow.
Isn't the point of arrays to be efficient? If you're just iterating through values, I think a cursor on a temporary (or permanent) table makes more sense than seeking commas, no? Also cleaner. Lookup "mysql DECLARE CURSOR".
For random access a temporary table with numerically indexed primary key. Unfortunately the fastest access you'll get is a hash table, not true random access.
Another way to see the same problem.
Hope helpfull
DELIMITER $$
CREATE PROCEDURE ARR(v_value VARCHAR(100))
BEGIN
DECLARE v_tam VARCHAR(100);
DECLARE v_pos VARCHAR(100);
CREATE TEMPORARY TABLE IF NOT EXISTS split (split VARCHAR(50));
SET v_tam = (SELECT (LENGTH(v_value) - LENGTH(REPLACE(v_value,',',''))));
SET v_pos = 1;
WHILE (v_tam >= v_pos)
DO
INSERT INTO split
SELECT SUBSTRING_INDEX(SUBSTRING_INDEX(v_value,',',v_pos),',', -1);
SET v_pos = v_pos + 1;
END WHILE;
SELECT * FROM split;
DROP TEMPORARY TABLE split;
END$$
CALL ARR('1006212,1006404,1003404,1006505,444,');
If we have one table like that
mysql> select * from user_mail;
+------------+-------+
| email | user |
+------------+-------+-
| email1#gmail | 1 |
| email2#gmail | 2 |
+------------+-------+--------+------------+
and the array table:
mysql> select * from user_mail_array;
+------------+-------+-------------+
| email | user | preferences |
+------------+-------+-------------+
| email1#gmail | 1 | 1 |
| email1#gmail | 1 | 2 |
| email1#gmail | 1 | 3 |
| email1#gmail | 1 | 4 |
| email2#gmail | 2 | 5 |
| email2#gmail | 2 | 6 |
We can select the rows of the second table as one array with CONCAT function:
mysql> SELECT t1.*, GROUP_CONCAT(t2.preferences) AS preferences
FROM user_mail t1,user_mail_array t2
where t1.email=t2.email and t1.user=t2.user
GROUP BY t1.email,t1.user;
+------------+-------+--------+------------+-------------+
| email | user | preferences |
+------------+-------+--------+------------+-------------+
|email1#gmail | 1 | 1,3,2,4 |
|email2#gmail | 2 | 5,6 |
+------------+-------+--------+------------+-------------+
In MYSQL version after 5.7.x, you can use JSON type to store an array. You can get value of an array by a key via MYSQL.
Inspired by the function ELT(index number, string1, string2, string3,…),I think the following example works as an array example:
set #i := 1;
while #i <= 3
do
insert into table(val) values (ELT(#i ,'val1','val2','val3'...));
set #i = #i + 1;
end while;
Hope it help.
Here is an example for MySQL for looping through a comma delimited string.
DECLARE v_delimited_string_access_index INT;
DECLARE v_delimited_string_access_value VARCHAR(255);
DECLARE v_can_still_find_values_in_delimited_string BOOLEAN;
SET v_can_still_find_values_in_delimited_string = true;
SET v_delimited_string_access_index = 0;
WHILE (v_can_still_find_values_in_delimited_string) DO
SET v_delimited_string_access_value = get_from_delimiter_split_string(in_array, ',', v_delimited_string_access_index); -- get value from string
SET v_delimited_string_access_index = v_delimited_string_access_index + 1;
IF (v_delimited_string_access_value = '') THEN
SET v_can_still_find_values_in_delimited_string = false; -- no value at this index, stop looping
ELSE
-- DO WHAT YOU WANT WITH v_delimited_string_access_value HERE
END IF;
END WHILE;
this uses the get_from_delimiter_split_string function defined here: https://stackoverflow.com/a/59666211/3068233
I Think I can improve on this answer. Try this:
The parameter 'Pranks' is a CSV. ie. '1,2,3,4.....etc'
CREATE PROCEDURE AddRanks(
IN Pranks TEXT
)
BEGIN
DECLARE VCounter INTEGER;
DECLARE VStringToAdd VARCHAR(50);
SET VCounter = 0;
START TRANSACTION;
REPEAT
SET VStringToAdd = (SELECT TRIM(SUBSTRING_INDEX(Pranks, ',', 1)));
SET Pranks = (SELECT RIGHT(Pranks, TRIM(LENGTH(Pranks) - LENGTH(SUBSTRING_INDEX(Pranks, ',', 1))-1)));
INSERT INTO tbl_rank_names(rank)
VALUES(VStringToAdd);
SET VCounter = VCounter + 1;
UNTIL (Pranks = '')
END REPEAT;
SELECT VCounter AS 'Records added';
COMMIT;
END;
This method makes the searched string of CSV values progressively shorter with each iteration of the loop, which I believe would be better for optimization.
I would try something like this for multiple collections. I'm a MySQL beginner. Sorry about the function names, couldn't decide on what names would be best.
delimiter //
drop procedure init_
//
create procedure init_()
begin
CREATE TEMPORARY TABLE if not exists
val_store(
realm varchar(30)
, id varchar(30)
, val varchar(255)
, primary key ( realm , id )
);
end;
//
drop function if exists get_
//
create function get_( p_realm varchar(30) , p_id varchar(30) )
returns varchar(255)
reads sql data
begin
declare ret_val varchar(255);
declare continue handler for 1146 set ret_val = null;
select val into ret_val from val_store where id = p_id;
return ret_val;
end;
//
drop procedure if exists set_
//
create procedure set_( p_realm varchar(30) , p_id varchar(30) , p_val varchar(255) )
begin
call init_();
insert into val_store (realm,id,val) values (p_realm , p_id , p_val) on duplicate key update val = p_val;
end;
//
drop procedure if exists remove_
//
create procedure remove_( p_realm varchar(30) , p_id varchar(30) )
begin
call init_();
delete from val_store where realm = p_realm and id = p_id;
end;
//
drop procedure if exists erase_
//
create procedure erase_( p_realm varchar(30) )
begin
call init_();
delete from val_store where realm = p_realm;
end;
//
call set_('my_array_table_name','my_key','my_value');
select get_('my_array_table_name','my_key');
Rather than Saving data as a array or in one row only you should be making diffrent rows for every value received. This will make it much simpler to understand rather than putting all together.
Have you tried using PHP's serialize()?
That allows you to store the contents of a variable's array in a string PHP understands and is safe for the database (assuming you've escaped it first).
$array = array(
1 => 'some data',
2 => 'some more'
);
//Assuming you're already connected to the database
$sql = sprintf("INSERT INTO `yourTable` (`rowID`, `rowContent`) VALUES (NULL, '%s')"
, serialize(mysql_real_escape_string($array, $dbConnection)));
mysql_query($sql, $dbConnection) or die(mysql_error());
You can also do the exact same without a numbered array
$array2 = array(
'something' => 'something else'
);
or
$array3 = array(
'somethingNew'
);
I'm looking to find records in a table that match a specific number that the user enters. So, the user may enter 12345, but this could be 123zz4-5 in the database.
I imagine something like this would work, if PHP functions worked in MySQL.
SELECT * FROM foo WHERE preg_replace("/[^0-9]/","",bar) = '12345'
What's the equivalent function or way to do this with just MySQL?
Speed is not important.
I realise that this is an ancient topic but upon googling this problem I couldn't find a simple solution (I saw the venerable agents but think this is a simpler solution) so here's a function I wrote, seems to work quite well.
DROP FUNCTION IF EXISTS STRIP_NON_DIGIT;
DELIMITER $$
CREATE FUNCTION STRIP_NON_DIGIT(input VARCHAR(255))
RETURNS VARCHAR(255)
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
WHILE iterator < (LENGTH(input) + 1) DO
IF SUBSTRING(input, iterator, 1) IN ( '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ) THEN
SET output = CONCAT(output, SUBSTRING(input, iterator, 1));
END IF;
SET iterator = iterator + 1;
END WHILE;
RETURN output;
END
$$
You can easily do what you want with REGEXP_REPLACE (compatible with MySQL 8+ and MariaDB 10.0.5+)
REGEXP_REPLACE(expr, pat, repl[, pos[, occurrence[, match_type]]])
Replaces occurrences in the string expr that match the regular expression specified by the pattern pat with the replacement string repl, and returns the resulting string. If expr, pat, or repl is NULL, the return value is NULL.
Go to REGEXP_REPLACE doc: MySQL or MariaDB
Try it:
SELECT REGEXP_REPLACE('123asd12333', '[a-zA-Z]+', '');
Output:
12312333
Updated 2022: as per Marlom's answer, you can now use REGEX_REPLACE - which will perform even better than my historical answer here.
Most upvoted answer above isn't the fastest.
Full kudos to them for giving a working proposal to bounce off!
This is an improved version:
DELIMITER ;;
DROP FUNCTION IF EXISTS `STRIP_NON_DIGIT`;;
CREATE DEFINER=`root`#`localhost` FUNCTION `STRIP_NON_DIGIT`(input VARCHAR(255)) RETURNS VARCHAR(255) CHARSET utf8
READS SQL DATA
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
DECLARE lastDigit INT DEFAULT 1;
DECLARE len INT;
SET len = LENGTH(input) + 1;
WHILE iterator < len DO
-- skip past all digits
SET lastDigit = iterator;
WHILE ORD(SUBSTRING(input, iterator, 1)) BETWEEN 48 AND 57 AND iterator < len DO
SET iterator = iterator + 1;
END WHILE;
IF iterator != lastDigit THEN
SET output = CONCAT(output, SUBSTRING(input, lastDigit, iterator - lastDigit));
END IF;
WHILE ORD(SUBSTRING(input, iterator, 1)) NOT BETWEEN 48 AND 57 AND iterator < len DO
SET iterator = iterator + 1;
END WHILE;
END WHILE;
RETURN output;
END;;
Testing 5000 times on a test server:
-- original
Execution Time : 7.389 sec
Execution Time : 7.257 sec
Execution Time : 7.506 sec
-- ORD between not string IN
Execution Time : 4.031 sec
-- With less substrings
Execution Time : 3.243 sec
Execution Time : 3.415 sec
Execution Time : 2.848 sec
While it's not pretty and it shows results that don't match, this helps:
SELECT * FROM foo WHERE bar LIKE = '%1%2%3%4%5%'
I would still like to find a better solution similar to the item in the original question.
There's no regexp replace, only a plain string REPLACE().
MySQL has the REGEXP operator, but it's only a match tester not a replacer, so you would have to turn the logic inside-out:
SELECT * FROM foo WHERE bar REGEXP '[^0-9]*1[^0-9]*2[^0-9]*3[^0-9]*4[^0-9]*5[^0-9]*';
This is like your version with LIKE but matches more accurately. Both will perform equally badly, needing a full table scan without indexes.
On MySQL 8.0+ there is a new native function called REGEXP_REPLACE. A clean solution to this question would be:
SELECT * FROM foo WHERE REGEXP_REPLACE(bar,'[^0-9]+',"") = '12345'
The simplest way I can think to do it is to use the MySQL REGEXP operator a la:
WHERE foo LIKE '1\D*2\D*3\D*4\D*5'
It's not especially pretty but MySQL doesn't have a preg_replace function so I think it's the best you're going to get.
Personally, if this only-numeric data is so important, I'd keep a separate field just to contain the stripped data. It'll make your lookups a lot faster than with the regular expression search.
This blog post details how to strip non-numeric characters from a string via a MySQL function:
SELECT NumericOnly("asdf11asf");
returns 11
http://venerableagents.wordpress.com/2011/01/29/mysql-numeric-functions/
There's no regex replace as far as I'm concerned, but I found this solution;
--Create a table with numbers
DROP TABLE IF EXISTS ints;
CREATE TABLE ints (i INT UNSIGNED NOT NULL PRIMARY KEY);
INSERT INTO ints (i) VALUES
( 1), ( 2), ( 3), ( 4), ( 5), ( 6), ( 7), ( 8), ( 9), (10),
(11), (12), (13), (14), (15), (16), (17), (18), (19), (20);
--Then extract the numbers from the specified column
SELECT
bar,
GROUP_CONCAT(SUBSTRING(bar, i, 1) ORDER BY i SEPARATOR '')
FROM foo
JOIN ints ON i BETWEEN 1 AND LENGTH(bar)
WHERE
SUBSTRING(bar, i, 1) IN ('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')
GROUP BY bar;
It works for me and I use MySQL 5.0
Also I found this place that could help.
I have a similar situation, matching products to barcodes where the barcode doesn't store none alpha numerics sometimes, so 102.2234 in the DB needs to be found when searching for 1022234.
In the end I just added a new field, reference_number to the products tables, and have php strip out the none alpha numerics in the product_number to populate reference_number whenever a new products is added.
You'd need to do a one time scan of the table to create all the reference_number fields for existing products.
You can then setup your index, even if speed is not a factor for this operation, it is still a good idea to keep the database running well so this query doesn't bog it down and slow down other queries.
I came across this solution. The top answer by user1467716 will work in phpMyAdmin with a small change: add a second delimiter tag to the end of the code.
phpMyAdmin version is 4.1.14; MySQL version 5.6.20
I also added a length limiter using
DECLARE count INT DEFAULT 0; in the declarations
AND count < 5 in the WHILE statement
SET COUNT=COUNT+1; in the IF statement
Final form:
DROP FUNCTION IF EXISTS STRIP_NON_DIGIT;
DELIMITER $$
CREATE FUNCTION STRIP_NON_DIGIT(input VARCHAR(255))
RETURNS VARCHAR(255)
BEGIN
DECLARE output VARCHAR(255) DEFAULT '';
DECLARE iterator INT DEFAULT 1;
DECLARE count INT DEFAULT 0;
WHILE iterator < (LENGTH(input) + 1) AND count < 5 DO --limits to 5 chars
IF SUBSTRING(input, iterator, 1) IN ( '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ) THEN
SET output = CONCAT(output, SUBSTRING(input, iterator, 1));
SET COUNT=COUNT+1;
END IF;
SET iterator = iterator + 1;
END WHILE;
RETURN output;
END
$$
DELIMITER $$ --added this
How big is table with foo? If it is small, and speed really doesn't matter, you might pull the row ID and foo, loop over it using the PHP replace functions to compare, and then pull the info you want by row number.
Of course, if the table is too big, this won't work well.
try this example. this is used for phone numbers, however you can modify it for your needs.
-- function removes non numberic characters from input
-- returne only the numbers in the string
CREATE DEFINER =`root`#`localhost` FUNCTION `remove_alpha`(inputPhoneNumber VARCHAR(50))
RETURNS VARCHAR(50)
CHARSET latin1
DETERMINISTIC
BEGIN
DECLARE inputLenght INT DEFAULT 0;
-- var for our iteration
DECLARE counter INT DEFAULT 1;
-- if null is passed, we still return an tempty string
DECLARE sanitizedText VARCHAR(50) DEFAULT '';
-- holder of each character during the iteration
DECLARE oneChar VARCHAR(1) DEFAULT '';
-- we'll process only if it is not null.
IF NOT ISNULL(inputPhoneNumber)
THEN
SET inputLenght = LENGTH(inputPhoneNumber);
WHILE counter <= inputLenght DO
SET oneChar = SUBSTRING(inputPhoneNumber, counter, 1);
IF (oneChar REGEXP ('^[0-9]+$'))
THEN
SET sanitizedText = Concat(sanitizedText, oneChar);
END IF;
SET counter = counter + 1;
END WHILE;
END IF;
RETURN sanitizedText;
END
to use this user defined function (UDF).
let's say you have a column of phone numbers:
col1
(513)983-3983
1-838-338-9898
phone983-889-8383
select remove_alpha(col1) from mytable
The result would be;
5139833983
18383389898
9838898383
thought I would share this since I built it off the function from here. I rearranged just so I can read it easier (I'm just server side).
You call it by passing in a table name and column name to have it strip all existing non-numeric characters from that column. I inherited a lot of bad table structures that put a ton of int fields as varchar so I needed a way to clean these up quickly before I can modify the column to an integer.
drop procedure if exists strip_non_numeric_characters;
DELIMITER ;;
CREATE PROCEDURE `strip_non_numeric_characters`(
tablename varchar(100)
,columnname varchar(100)
)
BEGIN
-- =============================================
-- Author: <Author,,David Melton>
-- Create date: <Create Date,,2/26/2019>
-- Description: <Description,,loops through data and strips out the bad characters in whatever table and column you pass it>
-- =============================================
#this idea was generated from the idea STRIP_NON_DIGIT function
#https://stackoverflow.com/questions/287105/mysql-strip-non-numeric-characters-to-compare
declare input,output varchar(255);
declare iterator,lastDigit,len,counter int;
declare date_updated varchar(100);
select column_name
into date_updated
from information_schema.columns
where table_schema = database()
and extra rlike 'on update CURRENT_TIMESTAMP'
and table_name = tablename
limit 1;
#only goes up to 255 so people don't run this for a longtext field
#just to be careful, i've excluded columns that are part of keys, that could potentially mess something else up
set #find_column_length =
concat("select character_maximum_length
into #len
from information_schema.columns
where table_schema = '",database(),"'
and column_name = '",columnname,"'
and table_name = '",tablename,"'
and length(ifnull(character_maximum_length,100)) < 255
and data_type in ('char','varchar')
and column_key = '';");
prepare stmt from #find_column_length;
execute stmt;
deallocate prepare stmt;
set counter = 1;
set len = #len;
while counter <= ifnull(len,1) DO
#this just removes it by putting all the characters before and after the character i'm looking at
#you have to start at the end of the field otherwise the lengths don't stay in order and you have to run it multiple times
set #update_query =
concat("update `",tablename,"`
set `",columnname,"` = concat(substring(`",columnname,"`,1,",len - counter,"),SUBSTRING(`",columnname,"`,",len - counter,",",counter - 1,"))
",if(date_updated is not null,concat(",`",date_updated,"` = `",date_updated,"`
"),''),
"where SUBSTRING(`",columnname,"`,",len - counter,", 1) not REGEXP '^[0-9]+$';");
prepare stmt from #update_query;
execute stmt;
deallocate prepare stmt;
set counter = counter + 1;
end while;
END ;;
DELIMITER ;
To search for numbers that match a particular numeric pattern in a string, first remove all the alphabets and special characters in a similar manner as below then convert the value to an integer and then search
SELECT *
FROM foo
WHERE Convert(Regexp_replace(bar, '[a-zA-Z]+', ''), signed) = 12345
I think you don't need complicated functions for that.
I've found a REGEXP_REPLACE solution using built-in mysql character class names. You can read about them in a table in the docs. Basically, they are mysql-specific names for commonly matched groups of characters like [:alnum:] for alpha-numeric characters, [:alpha:] for only alphabetic characters and so on.
So my version of the REGEXP_REPLACE:
REGEXP_REPLACE('My number is: +59 (29) 889-23-56', '[[:alpha:][:blank:][:punct:][:cntrl:]]', '')
will yield 59298892356 as per requirements.
Being someone that has version 5.7, doesn't have the privilege to create functions, and it's not practical to bring my data down into my code, I found Nelson Miranda's Answer amazing. I wanted to share it in a subquery version which I found more useful.
DROP TABLE IF EXISTS ints;
CREATE TABLE ints (i INT UNSIGNED NOT NULL PRIMARY KEY);
INSERT INTO ints (i) VALUES
( 1), ( 2), ( 3), ( 4), ( 5), ( 6), ( 7), ( 8), ( 9), (10),
(11), (12), (13), (14), (15), (16), (17), (18), (19), (20);
SELECT * FROM foo f
WHERE (SELECT GROUP_CONCAT(SUBSTRING(f.bar, i, 1) ORDER BY i SEPARATOR '')
FROM ints
WHERE i BETWEEN 1 AND LENGTH(f.bar)
AND SUBSTRING(f.bar, i, 1) IN ('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')) = '12345'
Note: The ints table can be a temporary table.
If you are on MySQL 5.7 or below, and you just need something quick and dirty without having to define a new function, and you have a small number of known non-numerics to filter out, something like this can work well...
Select replace(replace(replace(replace(replace(phone, '+1', ''), '(', ''), ')', ''), '-', ''), ' ', '') from customers;