Home

Awesome

Description

Squawk is a library and command line tool for running SQL queries against structured/semi-structured static files. (e.g. Apache logs, csv files, tcpdump output).

License

BSD

See LICENSE

Goal

The purpose is Squawk is to make querying for data in log files or other structured files easier. Everything that Squawk does can be done by combining various unix tools, but Squawk makes it ever easier to express more complex relationships. It is in no way a database or meant to be used as such. It's merely a reporting tool.

Squawk can be used from the command line for ad-hoc queries, and it can also be used as a library as a part of a more in-depth reporting tool.

Status

Still in major development. API is guaranteed to change.

Requirements

Supported SQL Features

Departures from Standard SQL

Parsers

Output Formats

Examples

SQL query on the command line::

$ squawk "SELECT COUNT(1) AS n, status FROM access.log GROUP BY status ORDER BY n DESC"
n	| status
----------------------------------------
381353	| 200
180668	| 302
17976	| 404
12952	| 301
10836	| 304
735	| 403
420	| 206
376	| 416
123	| 400
46	| 500
5	| 502
3	| 408
3	| 405
1	| 504

SQL based query through API::

query = Query(
    "SELECT COUNT(1) AS n, remote_addr"
    " FROM file"
    " WHERE status = 200"
    "  AND remote_addr != '-'"
    " GROUP BY remote_addr"
    " ORDER BY n DESC"
    " LIMIT 10")
source = AccessLogParser("access.log")
output_console(query(source))

# or

query = Query(
    "SELECT COUNT(1) AS n, remote_addr"
    " FROM file"
    " WHERE status = 200"
    "  AND remote_addr != '-'"
    " GROUP BY remote_addr"
    " ORDER BY n DESC"
    " LIMIT 10")
source = AccessLogParser("access.log")
for row in query(source):
    print row

Code generated query::

source = AccessLogParser("access.log")
filtered = Filter(source, lambda row:row['status'] == 200)
group_by = GroupBy(filtered, group_by=["remote_addr"], columns=[
    lambda:Column('remote_addr'),
    lambda:CountAggregate(None, 'count(1)')])
order_by = OrderBy(group_by, 'count(1)', True)
limit = LimitOffset(order_by, 10)
for row in limit:
    print row