13

I am inserting a list of dictionaries to a PostgreSQL database. The list will be growing quickly and the number of dict values (columns) is around 30. The simplified data:

projects = [
{'name': 'project alpha', 'code': 12, 'active': True},
{'name': 'project beta', 'code': 25, 'active': True},
{'name': 'project charlie', 'code': 46, 'active': False}
]

Inserting the data into the PostgreSQL database with the following code does work (as in this answer), but I am worried about executing too many queries.

for project in projects:
    columns = project.keys()
    values = project.values()

    query = """INSERT INTO projects (%s) VALUES %s;"""

    # print(cursor.mogrify(query, (AsIs(','.join(project.keys())), tuple(project.values()))))

    cursor.execute(query, (AsIs(','.join(columns)), tuple(values)))

conn.commit()

Is there a better practice? Thank you so much in advance for your help!

PythonSherpa
  • 2,560
  • 3
  • 19
  • 40
  • 1
    You can use copy statement. https://stackoverflow.com/questions/2271787/psycopg2-postgresql-python-fastest-way-to-bulk-insert – Nurullah Macun Mar 01 '19 at 14:41

3 Answers3

17

Use execute_values() to insert many rows in a single query.

import psycopg2
from psycopg2.extras import execute_values

# ...

projects = [
{'name': 'project alpha', 'code': 12, 'active': True},
{'name': 'project beta', 'code': 25, 'active': True},
{'name': 'project charlie', 'code': 46, 'active': False}
]

columns = projects[0].keys()
query = "INSERT INTO projects ({}) VALUES %s".format(','.join(columns))

# convert projects values to list of lists
values = [[value for value in project.values()] for project in projects]

execute_values(cursor, query, values)
conn.commit()
klin
  • 112,967
  • 15
  • 204
  • 232
3

Another performant option that does not require so much data munging for a list of dictionaries is execute_batch (new in psycopg2 version 2.7).

For example:

import psycopg2
from psycopg2.extras import execute_batch

values = [{'name': 'project alpha', 'code': 12, 'active': True}, ...]
query = "INSERT INTO projects VALUES (%(name)s, %(code)s, %(active)s)"

execute_batch(cursor, query, values)
conn.commit()

https://www.psycopg.org/docs/extras.html#psycopg2.extras.execute_batch

bensentropy
  • 1,293
  • 13
  • 17
2

You can use bulk loading to make it faster.

https://www.citusdata.com/blog/2017/11/08/faster-bulk-loading-in-postgresql-with-copy/

Johnny Chia
  • 92
  • 5