String over columns

From SQLZOO
Jump to navigation Jump to search

Normally a string can be searched for across many columns using the OR function but this is not efficient as this could lead to errors while using a CONCAT function ( || for oracle, + for sqlserver) could do the same task but can get rid of the risk of errors

schema:scott
DROP TABLE bedrooms
 CREATE TABLE bedrooms(
  name VARCHAR(20),
  floorcolor VARCHAR(20),
  ceilingcolor VARCHAR(20),
  wallcolor VARCHAR(20));
INSERT INTO bedrooms VALUES ('Jim','RED','GREEN','YELLOW');
INSERT INTO bedrooms VALUES ('Bob','YELLOW','BLUE','BLACK');
INSERT INTO bedrooms VALUES ('Allan','BLUE','PINK','BLACK');
INSERT INTO bedrooms VALUES ('George','BLUE','GREEN','OAK');
SELECT name FROM bedrooms
 WHERE floorcolor = 'YELLOW'
    OR ceilingcolor = 'YELLOW'
    OR wallcolor = 'YELLOW'

Using OR increases the chances for careless mistakes instead CONCAT could be used to make the query more efficient.

To avoid problems make sure to add in separators and if a value can be null make sure to use it is wrapped in COALESCE or NVL for example:

COALESCE(floorcolor,' ').

SELECT name FROM bedrooms
 WHERE ':' || floorcolor || ':' || ceilingcolor || ':' || wallcolor || ':'
  LIKE '%:YELLOW:%'
SELECT name FROM bedrooms
 WHERE ':' + floorcolor + ':' + ceilingcolor + ':' + wallcolor + ':'
  LIKE '%:YELLOW:%'
SELECT name FROM bedrooms
 WHERE CONCAT (':',floorcolor,':',ceilingcolor,':',wallcolor,':')
  LIKE '%:YELLOW:%'
DataWars, Data Science Practice Projects - LogoDataWars: Practice Data Science/Analysis with +100 Real Life Projects