README.rst

queryparser

Tool for parsing and processing MySQL and ADQL SELECT queries

Designed to be used in conjunction with django-daiquri as a query processing backend but it can be easily used as a stand-alone tool or integrated into another project.

Installation

The easiest way to install the package is by using the pip tool:

pip install queryparser-python3

or if you are using an older version (2.7) of python

pip install queryparser-python2

Alternatively, you can clone the repository and install it from there. However, this step also requires generating the parser which is a slighly more elaborate process (see below).

Generating the parser from the git repository

To generate the parsers you need python (either 2 or 3), java above version 7, and antlr4 (antlr-4.*-complete.jar has to be installed inside the /usr/local/lib/ or /usr/local/bin/ directories).

After cloning the project run

make

and a lib directory will be created with the complete source for python2 and python3. After that run

python setup.py install

to install the generated parser in your virtual environment.

Parsing MySQL

Parsing and processing of MySQL queries can be done by creating an instance of the MySQLQueryProcessor class

from queryparser.mysql import MySQLQueryProcessor

qp = MySQLQueryProcessor()

feeding it a MySQL query

sql = "SELECT a FROM db.tab;"
qp.set_query(sql)

and running it with

qp.process_query()

After the processing, the processor object qp will include tables, columns, functions, and keywords used in the query or will raise a QuerySyntaxError if there are any syntax errors in the query.

Alternatively, passing the query at initialization automatically processes it.

Translating ADQL

Translation of ADQL queries is done similarly by first creating an instance of the ADQLQueryTranslator class

from queryparser.adql import ADQLQueryTranslator

adql = "SELECT TOP 100 POINT('ICRS', ra, de) FROM db.tab;"
adt = ADQLQueryTranslator(adql)

and calling

adt.to_mysql()

which returns a translated string representing a valid MySQL query if the ADQL query had no errors. The MySQL query can then be parsed with the MySQLQueryProcessor in the same way as shown above.

Testing

First, install pytest

pip install pytest

then run the test suite for a version of python you would like to test with

pytest lib/python2
pytest lib/python3