csv2pg 2.2.0

A simple and fast cli application to load a csv into postgres

csv2pg

A simple and fast cli application to load a csv into postgres

• low memory footprint
• blazing fast (pg copy)
• lines and fields validation (--skip-error)

Installation

https://pypi.org/project/csv2pg/

pip install --user csv2pg

Usage

$ csv2pg --help
Usage: csv2pg [OPTIONS] TABLE FILEPATH

  COPY FROM 'csv' TO 'postgres'

Options:
  -h, --host TEXT             database server host  [default: localhost]
  -p, --port INTEGER          database server port  [default: 5432]
  -d, --dbname TEXT           database user name  [default: $USER]
  -U, --username TEXT         database name to connect to  [default: $USER]
  -W, --password              force password prompt
  -v, --verbose
  --progress                  display progress bar
  --skip-error                detect, ignore and export errors to
                              <filepath>.err  [default: False]

  --header / --no-header      [default: True]
  --rownum / --no-rownum      include line number in a _rownum column
                              [default: False]

  --filename / --no-filename  include filename in a _filename column
                              [default: False]

  --delimiter TEXT            char separating the fields  [default: ,]
  --quotechar TEXT            char used to quote a field  [default: "]
  --doublequote               When True, escapechar is replaced by doubling
                              the quote char  [default: False]

  --escapechar TEXT           char used to esapce the quote char  [default: \]
  --lineterminator TEXT       line ending sequence  [default:  ]
  --null TEXT                 will be treated as NULL by postgres  [default: ]
  --encoding TEXT             [default: utf-8]
  --overwrite                 destroy table before inserting csv  [default:
                              False]

  --unlogged                  insert in an UNLOGGED table (faster)  [default:
                              False]

  --buffer INTEGER            size of the read buffer to be used by COPY FROM
                              [default: 8192]

  --version                   Show the version and exit.
  --help                      Show this message and exit.

Basic usage:

csv2pg -h localhost -p 5432 -U postgres -d postgres public.data data.csv --verbose

Basic usage with postgres environment variables:

PGHOST=localhost PGPORT=5432 PGDATABASE=postgres PGUSER=postgres PGPASSWORD= csv2pg public.data data.csv --verbose

Loading a tab delimited latin-1 encoded file in an unlogged table with _filename and _rownum columns, skipping errors and displaying progress bar:

PGPASSWORD= csv2pg -h localhost -p 25432 -d test -U test \
    --delimiter=$'\t' --encoding="iso-8859-1" \
    --overwrite --unlogged \
    --filename --rownum \
    --skip-error --progress \
public.data data.csv

Quick test

Start a postgres database:

docker run -d --rm \
    -p 25432:5432 \
    --name csv2pg-test \
    -e POSTGRES_DB=test \
    -e POSTGRES_USER=test \
    -e POSTGRES_PASSWORD=test \
postgres

Download and import a test asset file:

wget https://raw.githubusercontent.com/DavidLacroix/csv2pg/master/tests/assets/simple.csv .
PGPASSWORD=test csv2pg -h localhost -p 25432 -U test -d test public.data simple.csv --progress

From python

import csv2pg

HOST = "localhost"
PORT = 25432
DBNAME = "test"
USER = "test"
PASSWORD = "test"

csv2pg.copy_to(HOST, PORT, DBNAME, USER, PASSWORD, "public.data", "./simple.csv", verbose=True)

Precaution

• the --overwrite option will drop the table before inserting the new records in.
• the --rownum and --filename options will slightly increase the insertion time (increase the data to write on disk)
• the --skip-error option will slightly increase the insertion time (fields and lines validation)
• --verbose and --progress used together might spoil the console output