Deleting duplicate rows from a table

Question

I have a table in my database which has duplicate records that I want to delete. I don't want to create a new table with distinct entries for this. What I want is to delete duplicate entries from the existing table without the creation of any new table. Is there any way to do this?

 id           action
 L1_name      L1_data
 L2_name      L2_data
 L3_name      L3_data   
 L4_name      L4_data
 L5_name      L5_data
 L6_name      L6_data
 L7_name      L7_data
 L8_name      L8_data
 L9_name      L9_data
 L10_name     L10_data
 L11_name     L11_data
 L12_name     L12_data
 L13_name     L13_data 
 L14_name     L14_data
 L15_name     L15_data

see these all are my fields :
id is unique for every row.
L11_data is unique for respective action field.
L11_data is having company names while action is having name of the industries.

So in my data I'm having duplicate name of the companies in L11_data for their respective industries.

What I want is to have is unique name and other data of the companies in the particular industry stored in action. I hope I have stated my problem in a way that you people can understand it.

Answer 1

DELETE
FROM table_x a
WHERE rowid < ANY (
  SELECT rowid
  FROM table_x b
  WHERE a.someField = b.someField
   AND a.someOtherField = b.someOtherField
  )
WHERE (
  a.someField,
  a.someOtherField
  ) IN (
  SELECT c.someField,
   c.someOtherField
  FROM table_x c
  GROUP BY c.someField,
   c.someOtherField
  HAVING count(*) > 1
  )

In above query the combination of someField and someOtherField must identify the duplicates distinctively.

Answer 2

ALTER IGNORE TABLE 'table' ADD UNIQUE INDEX(your cols);

Duplicates get NULL, then you can delete them

Answer 3

Yes, assuming you have a unique ID field, you can delete all records that are the same except for the ID, but don't have "the minimum ID" for their group of values.

Example query:

DELETE FROM Table
WHERE ID NOT IN
(
SELECT MIN(ID)
FROM Table
GROUP BY Field1, Field2, Field3, ...
)

Notes:

• I freely chose "Table" and "ID" as representative names
• The list of fields ("Field1, Field2, ...") should include all fields except for the ID
• This may be a slow query depending on the number of fields and rows, however I expect it would be okay compared to alternatives

EDIT: In case you don't have a unique index, my recommendation is to simply add an auto-incremental unique index. Mainly because it's good design, but also because it will allow you to run the query above.