Skip to main content

Trivial Postgres Queue

Project description

Travis CI Status Code Coverage

Trivial Postgres Queue

This is a simple library that can place JSON blobs into a FIFO queue and later retrieve them.

The difference between this queue and other similar queues is that it utilizes FOR UPDATE SKIP LOCKED.

The advantage here is that queue items remain in the queue until the current transaction is committed. Upon rollback, the item is left in the queue untouched.

If other records are modified as part of a larger transaction, those changes are also rolled back. Making the larger queue processing operation atomic. Take the follow order of operations




In the above, none of the statements have any affect, and the queue item remains in the table to be “retried” by another consumer. Since FOR UPDATE is used, the queue item remains locked to avoid multiple consumers obtaining that item from the queue.

Python Library Usage

Database connection information can be provided via the library API.

import tpq

# Explicitly provide database connection information
q = tpq.Queue('queue_name', host='localhost', dbname='foobar')
q.put('{"foo": "bar"}')

# Or use shortcut functions:
tpq.put('queue_name', '{"foo": "bar"}', host='localhost', dbname='foobar')
tpq.get('queue_name', host='localhost', dbname='foobar')

# Or to take advantage of cooperative transactions, provide a connection:
q = tpq.Queue('queue_name', conn=connection)
q.put('{"foo": "bar"}')

# Which is also supported by shortcut functions:
tpq.put('queue_name', '{"foo": "bar"}', conn=connection)
tpq.get('queue_name', conn=connection)

Or, you can set the connection info in the environment:

$ # Export as URL
$ export TPQ_URL="postgresql://user:pass@localhost/foobar"

$ # Or separately
$ export TPQ_HOST=localhost
$ export TPQ_DB=foobar
$ export TPQ_USER=user
$ export TPQ_PASS=pass

Then omit the parameters:

import tpq

# Use an instance for multiple operations
with tpq.Queue('queue_name') as q:
    q.put('{"foo": "bar"}')
    data = q.get()

# Or use shortcut functions:
tpq.put('queue_name', '{"foo": "bar"}')


You can wait for an item to arrive using the wait argument.

import tpq

# Wait forever
tpq.get('queue_name', wait=0)

# Don't wait (also can omit the param).
tpq.get('queue_name', wait=-1)

# Wait specified number of seconds.
tpq.get('queue_name', wait=5)

Command Line Interface

Command line interface is also provided. JSON can be provided via a file or stdin (the default).

$ # Configure your database.
$ export TPQ_URL="postgresql://user:pass@localhost/foobar"

$ # JSON via stdin (default).
$ echo "{\"foo\": \"bar\"}" | tpq produce queue_name

$ # JSON via file.
$ tpq produce queue_name --file=message.json

$ # Explicitly provide JSON via stdin.
$ tpq produce queue_name --file=- < message.json

$ # Then read the item to stdout.
$ tpq consume queue_name
{'foo': 'bar'}

$ # If you have trouble (or for logging). Debug output goes to stderr.
$ TPQ_URL="postgresql://user:pass@localhost/foobar" tpq consume queue_name --debug
Read database config from environment
Parsing TPQ_URL
Database config found
Attempting to read item
Item read, returning
{'foo': 'bar'}

$ # You can wait on the CLI too...
$ # Forever:
$ tpq consume queue_name --wait=0

$ # Specified number of seconds:
$ tpq consume queue_name --wait=5

$ # The return code signals whether an item was received or not.
$ tpq consume queue_name --wait=-1
{'foo': 'bar'}
$ echo $?

# For an empty queue, you get 1
$ tpq consume queue_name --wait=-1
Queue empty
Traceback (most recent call last):
  File "/home/btimby/Code/tpq/tpq/", line 24, in consume
    print(get(opt['<name>'], wait=opt['--wait']))
  File "/home/btimby/Code/tpq/tpq/", line 266, in get
    return q.get(wait=wait)
  File "/home/btimby/Code/tpq/tpq/", line 233, in get
    raise QueueEmpty()
$ echo $?

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

tpq-1.7.tar.gz (8.7 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page