There are times when duplicate records somehow creep into a table despite your best efforts. This happens more in cases where data is loaded into table from other sources because during data loads, the integrity constraints are disabled.
The following SQL statement helps in deleting the duplicate rows in a table:
delete table_a where rowid not in (select min(rowid) from table_a group by column1, column2);
In the above Data Manipulation Language statement, the sub-query should be grouped by all the columns that are part of the primary/unique key. In the above case it is column1 and column2.