How to get number of affected rows in sqlalchemy?

ResultProxy objects have a rowcount property as well.

What Shiplu says is correct of course, however, to answer the question, in many cases you can easily make the matched columns equal the changed columns by including the condition that the value is different in the WHERE clause, i.e.:

UPDATE table
SET column_to_be_changed = "something different"
WHERE column_to_be_changed != "something different" AND [...your other conditions...]

rowcount will then return the number of affected rows, because it equals the number of matched rows.

Shiplu's analysis is 100% correct.

Pushing the discussion a bit further, here is how to display the updated rowcount and not the matched rowcount using sqlalchemy for MySQL: the engine needs to be created with the flag client_flag=0.

from sqlalchemy import create_engine
engine = create_engine(
    'mysql+pymysql://user:password@host:port/db', 
    connect_args={'client_flag':0}
)

To give a bit more details, the rowcount returned by MySQL depends on the CLIENT_FOUND_ROWS flag provided to the C-API function mysql_real_connect() as stated in MySQL documentation:

For UPDATE statements, the affected-rows value by default is the number of rows actually changed. If you specify the CLIENT_FOUND_ROWS flag to mysql_real_connect() when connecting to mysqld, the affected-rows value is the number of rows “found”; that is, matched by the WHERE clause.

The flag value is 2 (MySQL constants), and is added automatically by sqlalchemy when creating the engine as visible here.

The value of client_flag in connect_args allows to override this value.

Note: this might break sth from the sane_rowcount (only used in the ORM apparently) - in my case, I only use the Core of sqlalchemy. In last link:

# FOUND_ROWS must be set in CLIENT_FLAGS to enable
# supports_sane_rowcount.