Upload your data
Upload your data to at least one dataset in BigQuery (e.g.
YourData). You can also split your data across multiple BigQuery projects.
Create another dataset for the output of joins that you make between Google data and your data (e.g.
Give permission to Ads Data Hub service account
The name of your service account is in an email from Ads Data Hub support.
You must give the service account
dataViewer permission for the dataset containing your uploaded data (e.g.
YourData). If your data is split across multiple datasets, the service account will need
dataViewer permission for all of them.
The service account also requires
dataEditor permission for the dataset containing your data join outputs (e.g.
Write queries to join data
As a best practice for your uploaded data, identify each of your users with an encrypted ID or cookie. After setting up cookie matching, uploading your data, and giving the Ads Data Hub service account access, all tables containing Google, AdWords, or DoubleClick IDs will get new match tables, labeled
_match, that also contain your user IDs. For example the
adh.adwords_impressions table receives a match table called
adh.adwords_impressions_match containing your user IDs. Join a table with its aligning match table to join your data with Google data.
Mobile device IDs
If you upload a dataset containing resettable mobile device IDs, it also can also be joined to the
_rdid tables. For this to work, the device IDs in your data must be hashed with MD5. Device IDs in your data will then be joinable with device IDs in the
_rdid tables. To query
_rdid tables, add
_rdid to the name of any
adh. table that contains
user_id. As with any Ads Data Hub query, results will be aggregated to ensure user privacy.
Qualified table names, e.g.
projectname.datasetname.tablename, are required when joining your data with Google data.