Google Ads Reporting in Python

An AdWords API large scale reporting tool written in
Python. Reports are downloaded as plaintext files but connectivity with a database engine such as MySQL, PostgreSQL, or
MongoDB can be implemented upon request.

Overview

pyaw-reporting is an open-source Python package suitable for large scale
Google Ads reports. Output reports are comma-separated values (CSV) plaintext files. By default, the package uses 10
threads to download reports simultaneously from different accounts. The number of threads used can be modified using
the -n parameter. It has been successfully tested using +100 threads making it useful for heavy load AdWords Manager
Accounts.

Supported AdWords API version

The latest API version supported by this package is
v201809 with
googleads 28.0.0. Older versions of the API are not supported, nor the newer
Google Ads API.

Quick Start

Prerequisites

You will need Python 3.6+; this package installs googleads as a dependency.
Note that that support for Python 2.7
ended in 2019. Using a
virtual environment is recommended to avoid running the package with sudo.

An access token YAML file with the corresponding AdWords credentials is also necessary. By
default, the package will look for ~/googleads.yaml unless a different path is passed. The optional parameter
client_customer_id must be included in this YAML file, you should enter your AdWords Manager Account id
(formerly known as MCC id). This way, all accounts linked to the Manager Account will be retrieved. The sample file
example.yaml shows how your token should look like.

AWQL queries could be either passed in the command line
with the -a/--awql parameter or stored in a plaintext file and passed via the -q/--query parameter. The module
includes a sample query which will retrieve clicks and impressions per ad for the last 7 days.
Refer to Report Types and
The AdWords Query Language (AWQL) for more information
about these queries.

Installation

$ pip install pyaw-reporting

Usage

Command line

usage: awreporting [-h] (-a AWQL | -q QUERY_FILE) [-t TOKEN] [-o OUTPUT]
                   [-n NUM_THREAD] [-v]
AwReporting - Large scale AdWords reporting tool in Python
optional arguments:
  -h, --help                              show this help message and exit
  -t TOKEN, --token TOKEN                 specify AdWords YAML token path
  -o OUTPUT, --output OUTPUT              define output file name
  -n NUM_THREAD, --num-thread NUM_THREAD  set number of threads
  -v, --verbose                           display activity
required arguments:
  -a AWQL, --awql AWQL                    pass AWQL query
  -q QUERY_FILE, --query-file QUERY_FILE  ...or use a query file

For example:

$ awreporting -t example.yaml -a \
  "SELECT ExternalCustomerId, CampaignId, AdGroupId, Id, Date, Clicks, Impressions \
  FROM AD_PERFORMANCE_REPORT WHERE Impressions > 0 DURING LAST_7_DAYS" \
  -o adperformance.csv -n 100

Or, with a query file:

$ awreporting -t example.yaml -q query.txt -o adperformance.csv -n 100

If you experience problems downloading reports, check the resulting logs in awreporting.log or use the
-v/--verbose parameter to verify if something is wrong with your token or AWQL query.

Code

If you prefer embedding this package in your own code, you could do like so:

from awreporting import get_report
query = (
    "SELECT ExternalCustomerId, CampaignId, AdGroupId, Id, Date, Clicks, Impressions "
    "FROM AD_PERFORMANCE_REPORT "
    "WHERE Impressions > 0 "
    "DURING LAST_7_DAYS"
)
get_report('example.yaml', query, 'adperformance.csv', 100)

About the YAML token

The example token file provided example.yaml is not valid. Refer to
this guide if you are using the AdWords API for
the first time.

Disclaimer

This is neither an official AdWords API repository nor a clone of
AwReporting. Consider using
AwReporting if you are a Java developer. This framework no longer supports
Python 2.7; only Python 3.6 or greater are compatible since
googleads requires it.

Troubleshooting

We recommend that you try the app with a few number of threads first (the default is 10) and increase the amount
accordingly. The Google Ads server may complain when many API calls are made at the same time, but those exceptions are
handled by the app. We have successfully generated huge report files using 200 threads.

When using this tool it might be necessary to enable a DNS cache in your system, such as
nscd. This should eliminate DNS lookup problems when repeatedly
calling the AdWords API server. For example, if you find
many URLError: <urlopen error [Errno -2] Name or service not known> in your logs, enable the DNS cache.

In some Linux systems nscd is not enabled by default, but it can be started with:

# systemctl start nscd