.NET (2)
Database (71)
Delphi IDE (89)
Network (39)
Printing (3)
Strings (12)
VCL (83)
Windows with Delphi (280)
Exchange Links
About this site
| |
 |
Find duplicate rows in a table (no primary key)
22 comments. Current rating: 
(8 votes). Leave comments and/ or rate it.
Question:
I have a table of city names and need to find duplicate entries. The table does not have a primary key, so the duplicate rows are entirely identical.
Answer:
You can use the GROUP clause in your SELECT statement. Below are 3 different attempts. Try the first one first. It may not work with your DBMS but if it works, it's most convenient. The second one is an ok choice as it returns all rows but the duplicate rows in the 'See Also' box.
The last version returns all city names with a count value and you manually have to look for those with a value > 1. If you need to do this more often, you should consider creating a VIEW or a temporary table with this query and then do a SELECT on this VIEW/ temp table.
In Microsoft SQL-Server, you can use the HAVING clause, as shown at the bottom.
And: If you do have a primary key defined, see the other tip mentioned at the top. This would work in that situation, but using the unique ID is better.
 |  | |
// Return all duplicate cities and how often they appear // Works with ORACLE select city_name, count(city_name) as cnt from areas group by city_name where cnt>1 // not all SQL dbms will support the reference to the count column cnt in the where clause. // The following will return ALL rows with counter, but sorted by number of appearances // Your duplicates will be at the top. // Works with MYSQL select city_name, count(city_name) as cnt from areas group by city_name order by cnt desc // finally, no back reference to count column cnt at all- // the following will work on all SQL dbms: // Return all cities and how often they appear select city_name, count(city_name) as cnt from areas group by city_name // version for Micrsoft's MSSQL Server // make use of the HAVING clause select city_name from areas group by city_name having count(*) > 1 | ||
 |  |
|
||||||||
Comments:
| You are on page 1 of 2, other pages: [1] 2 | ||
| ||
| ||
|
it's better
SELECT A.ID, B.ID FROM T1 A INNER JOIN T1 B ON (A.F1=B.F1 AND A.F2 AND B.F2 AND ...) WHERE A.ID<B.ID | ||
| ||
| ||
|
When you want to find duplicate combination of multiple columns,
group by is the best approach. SELECT col1, col2,col3,count(col3)--put the last column of duplicate combination in Count() FROM table1 GROUP BY col1,col2,col3 Here as col1,col2,col3 combination will be unique as they are grouped by GROUP BY clause and count() returns the count. We can add a HAVING Clause to filter combinations for whom the count() is > 1 SELECT col1, col2,col3,count(col3)--put the last column of duplicate combination in Count() FROM table1 GROUP BY col1,col2,col3 HAVING COUNT(col3) >1 | ||
| ||
|
Use full
| ||
| ||
|
used it in MySQL and it worked great. very helpful, thanks
| ||
| ||
|
I have a slightly different need.
I have a table that carries multiple rows for a single key. Example: Two cities in same state. (Cities Marlboro, Middleboro in State of Califernia) My need is to show one row per state and cancatenate cities in that row, e.g., Califernia, Marlboro and Middleboro Any suggestions on how to do this? | ||
| ||
|
I have a slightly different need.
I have a table that carries multiple rows for a single key. Example: Two cities in same state. (Cities Marlboro, Middleboro in State of Califernia) My need is to show one row per state and cancatenate cities in that row, e.g., Califernia, Marlboro and Middleboro Any suggestions on how to do this? Your Solution can be: SELECT state1 . * , state2.name FROM `state` state1 LEFT JOIN state state2 ON state1.statename = state2.statename WHERE state1.city <> state2.city GROUP BY state1.statename | ||
| ||
|
Good ,It worked for me too. thanks.
| ||
| ||
|
Can someone help me with my problem?
My table looks like this CID FNAME MIN MAX COM A OP1 0 23 5 A OP1 24 35 2 A OP1 36 99 1 A OP2 0 23 5 A OP2 24 35 2 A OP2 36 99 0 A OP3 0 23 5 A OP3 24 35 2 A OP3 36 99 1 B OP1 0 23 9 B OP1 24 99 2 B OP2 0 23 9 B OP2 24 99 2 B OP3 0 23 7 B OP3 24 35 3 B OP3 36 99 1 The expected duplicate results i need are: A OP1 A OP3 B OP1 B OP2 | ||
| You are on page 1 of 2, other pages: [1] 2 |
