Data Exploration Query Language (DXQL)

Requires Python 3.7

Usage

1. Import dxql.search.Pipeline into your project
2. Instantiate a Pipeline using Pipeline.create_pipeline(query-string)
3. Use the new pipeline to search over an iterable of dicts using pipeline.execute(events)

Example:

from dxql.search import Pipeline
pipeline = Pipeline.create_pipeline('search ip=192.168.1.10')
results = pipeline.execute(events)

events can be any iterable. To search a file, just pass the opened file to pipeline.execute(). Each line of the file will be considered an event.

Example:

# myfile.json is a file where each line is a JSON dictionary
with open('myfile.json') as file:
    results = pipeline.execute(file)

Searching

Searching is inspired by Splunk's query language.

Throughout the rest of this document, I will use the terms "search" and "query" interchangebly.

A query can consist of multiple commands separated by a pipe (|). Imagine a multiple-command search as a "pipeline" where each command is applied to the data in turn, with the data being fed from one command to the next until the end of the pipeline.

There are four commands available:

1. search

The search command allows you to filter the data using key-value pairs and modifiers like OR and NOT. It must be the first command in the query.

Usage:

search <expression>...

<expression>

<comparison-expression> | NOT <expression> | <expression> OR <expression>

<comparison-expression>

<field><operator><value>

<operator>

= | != | < | <= | > | >=

Examples

Retrieving data from an index

This search will return all data from the geoip index.

search index=geoip

Retrieving GeoIP data for specific IPs

Use the OR modifier to specify multiple values for a field.

search index=geoip ip=192.168.1.10 OR ip=192.168.1.11

Retrieving GeoIP data for all IPs except one

search index=geoip ip!=192.168.1.15

or

search index=geoip NOT ip=192.268.1.15

Retrieving data for a specific IP from multiple indices

It is not required to search by index.

search ip=192.168.1.15

The above search will return data with ip=192.168.1.15 from all indices (in this case, data from indices geoip and ip_rdap will be returned; events in rdap do not contain an ip field).

2. fields

The fields command allows you to display only the fields you want to see.

Usage

fields <field>...

Example

Remove all fields from the results except for ip and continent_name:

search index=geoip | fields ip continent_name

3. join

The join command allows you to join data together by a field (the "by-field"). Each event that shares the same value for the by-field is joined together under one event. This allows you to join data from two disparate data sources.

Usage

join BY <by-field>

Example

Join an IP with its associated RDAP data using the ip_rdap and rdap indices:

search index=ip_rdap OR index=rdap | join BY handle

handle is the 'by-field', the field that is shared by the different kinds of data.

4. prettyprint

The prettyprint command may only be used as the last command in the search. It allows you to print the result set in a prettier fashion than plain JSON blobs.

Usage

prettyprint format=<format>

<format>

json | table

Examples

Print results as pretty JSON

Using format=json still prints each result as JSON but with newlines and indentation.

search index=rdap | prettyprint format=json

Print results as a table

Using format=table prints the results as a formatted table.

search index=rdap | prettyprint format=table

If there are a lot of fields in the result set, the results will overflow onto the next line(s); therefore, it is recommended to pare down unwanted fields using fields before using prettyprint format=table. This happens expecially when joining ip_rdap and rdap data together. Many IPs share the same rdap data, so the IP values will become very long. I recommend specifying the IP(s) you are interested in before doing the join.