To parse the difference between replace into and insert in an SQL statement

This article is a detailed analysis of the difference between replace into and insert into SQL statements, the need for a friend under the reference

The run of replace is similar to insert. With one exception, if an old record in the table has the same value as a new record for primary key or a unique index, the old record is deleted before the new record is inserted.
Note that it is meaningless to use a Replace statement unless the table has a primary key or a unique index. The statement is the same as the insert, because no indexes are used to determine whether the new row replicates other rows.

The values for all columns are taken from the value specified in the Replace statement. All missing columns are set to their default values, which are the same as inserts. You cannot reference a value from the current row, nor can you use a value in a new row. If you use an assignment such as "SET col_name = col_name + 1", a reference to the column name on the right side is treated as default (col_name). Therefore, the assignment is equivalent to set col_name = DEFAULT (col_name) + 1.

In order to be able to use replace, you must have both insert and delete permissions for the table.

The Replace statement returns a number that indicates the number of rows affected. This number is the same as the number of rows that were deleted and inserted. If the number is 1 for a single line replace, the row is inserted and no rows are deleted. If the number is greater than 1, one or more old rows are deleted before the new row is inserted. If a table contains more than one unique index, and a new row duplicates a value from a different old row in a different unique index, it is possible that a single row replaces more than one old row.

The number of rows affected can easily determine whether replace has added only one row, or if replace has also replaced other rows: check to see if the number is 1 (added) or larger (replace).

If you are using the C API, you can use the Mysql_affected_rows () function to get the number of rows affected.

Currently, you cannot replace a table in a subquery and select from the same table.

The following is a detailed description of the algorithm (this algorithm is also used for load DATA ...). REPLACE):
1. Try inserting a new row into the table
2. When an insert fails because of a duplicate keyword error for a primary key or unique keyword:
A. Delete conflicting rows that contain duplicate key values from the table
B. Try inserting the new row into the table again

Use the following format:
REPLACE [low_priority | Delayed]
[Into] tbl_name [(Col_name,...)]
VALUES ({expr | DEFAULT},...), (...),...
or:
REPLACE [low_priority | Delayed]
[Into] Tbl_name
SET col_name={expr | DEFAULT}, ...
or:
REPLACE [low_priority | Delayed]
[Into] tbl_name [(Col_name,...)]
SELECT ...