The Query Tool is an interactive UI within the Platform, which you can use to build, test and run queries.

The Query Tool helps you explore the datasets you can reference, and to build a query using InfoSum's query language, IQL. If you want to, you can then take the finished query and run it using the Platform API.

As discussed in the IQL reference documentation, InfoSum Platform supports two types of queries: insight queries and activation queries. Unless previously agreed otherwise, activation queries are disabled by default as it's important to consider legal and regulatory requirements. If required, please get in touch with your InfoSum contact to discuss enabling this feature.

Alongside the slight difference in syntax, the use case and output greatly differ between insight and activation queries. Insight queries return aggregated, statistical results whereas activation queries return a list of identifiers that relate to specific individuals. In order to submit an activation query, you must have selected an output column during the data normalization phase.

When you use the Query Tool, the first thing to select is the query type, either Insight or Activation, then start writing the query.

Tip: If you don't see the Activation query type, you will need to contact support@infosum.com to enable this option.

You can click the Copy to console icon in the left pane to copy the name of a dataset, category, or category and property to the query text. 

To copy a previous query, click the Copy to console icon to reproduce the whole query in the query text.

Insight Queries

For example:    

SELECT
  Aggregate(
    ACME.Age {20,30,40,50},
    AutoSports."sports lover"
)

FROM ACME INTERSECT AutoSports

When you're ready, select SUBMIT and then you will see the anonymized statistics in interactive histograms, joyplots and pie graphs. You can also review the query quality and view the query in JSON format. If your query contains any errors, a warning message appears under the query.

The above example query returns aggregate bins and representations data for a total of 34,100 records found at the intersection of two datasets, as shown.

The Platform automatically works out the best key for matching based on the highest fill rate (Email in this example). Using Advanced Settings you can override the InfoSum Platform automatic key choice based on the highest fill rate, as shown in the example below. 

Resubmitting the example query with the override key MAID as the matching key returns different results, with the total matching records now shown as 17,500 (down from 34,100 for the platform selected key, Email). 

Activation Queries

To run an activation query, the activation bunker you are querying must contain an imported dataset. Two activation Bunkers cannot be included in a single query.

For example:

SELECT TestActivation.device_id 
FROM TestActivation INTERSECT (ACME INTERSECT IdentityX)

When building an activation query, the Platform automatically works out the best key for matching based on the highest fill rate (Cookie ID in this example), as shown below.

Using Advanced Settings you can override the InfoSum Platform automatic key choice based on the highest fill rate, as shown. 

Using the override key Email as the matching key builds an activation query with different results, with the total rows for the output column device_id  now 21,000 (down from 33,500 for the platform selected key, Cookie ID).

You can use the IQL syntax below to export multiple output columns from an activation dataset to a destination CSV file. Click here to see the maximum number of output columns you can export.

SELECT
   activationTable.{column1,column2} 

FROM activationTable

For example:

SELECT
MultiOutput.{email,"cookie id",alternate_email,"Alternate Vehicle Registration Number"}

FROM MultiOutput

You can add multiple columns of identical string values to every row in your campaign’s output CSV file. These string values can then be used by in-house/downstream systems to recognize each identifier in the campaign’s output CSV file. This is very useful for recognizing and grouping a list of identifiers into different audience segments across technology providers.

Select Add a column of string values to add one or more columns of identical string values to every row in your campaign’s output CSV file.  Here, you can specify the column name and value for each column (there is no limit to the number of columns you can add). The example adds the column names Segment_1 and Segement_2  and the column values value_1 and value_2 as string values in your output file. 

Select the Column Delimiter used to separate each output column in your destination CSV file, such as tab, colon, semi-colon, pipe, space or caret. The example uses comma as the column delimiter.

email,cookie id,HeaderOverride,Alternate Vehicle Registration Number,segment

Select a different delimiter as the Multi-value Column Delimiter - this delimiter is used to separate multiple values within single columns of your destination CSV file. The example uses pipe as the multi-value column delimiter.

Select Override queried column headers with new headers to rename column headers in your campaign (please contact your Customer Success Representative if override queried column headers are not available). The new column headers you type here replace the specified headers to override in your destination CSV file. Always include the name of the activation file when specifying the header to override. The example renames the column header MultiOutput.alternate_email to HeaderOverride. 

If you've built an activation query, name the campaign and select Create campaign. You will then be taken to the Activation tab, where you can view the size of the resulting dataset and choose to push the results to an external destination, such as a Google Cloud Storage bucket, an SFTP server or S3.

The output for the example activation is shown below as a destination CSV file and underneath in its raw data format. Columns are separated by a comma and values within the "email" multi-value key column are separated by a pipe.

Example destination CSV file:

Raw data format: