Backend Parser

Backend parser for infinity is introduced in version 1.0.0. Setting the parser to backend in your query editor will allow you to use features such as alerting, grafana expressions ,recorded queries, enterprise query caching and shared dashboards.

Support for backend parser is available for JSON from version 1.0.0. Support for backend parser is available for CSV/TSV/GraphQL from version 1.1.0. Support for backend parser is available for XML/HTML from version 1.2.0.

Root selector/Field selector

For JSON, data received from the server should be in array format. If not, then the array must be specified by the root selector. Root selector must be in the format specified by gjson.

Backend parser uses gjson style selectors and legacy/default/frontend parser uses lodash type selectors.

Computed fields

Once you have some fields already and you want to compute new field based on existing columns, you can use computed fields to do. This is something similar to the Add field from calculation -> Binary Operation transformation but enhanced with powerful expression language. For example price * qty gives the multiplication value of two columns names price and qty

Filter

If you want to filter data, you can use filter expressions. The expression should yield true or false. Example: price > 500, name != 'MacBook' && price != null, name IN ('MacBook','MacBook Air') and !(name IN ('MacBook','MacBook Air')) are valid filter expressions.

Summarize

Backend parser also supports summarizing the fields into a metric. For example, count(something) or max(some-other-thing) or mean([some other thing]) can be calculated from the array of data. You can also use numeric options such as sum,min,max,mean,first and last. You can also use expressions like sum(price) / count(id).

When specifying the field name, all special characters must be replaced with - and also transformed to lower case. Or the field name can also be specified within square brackets if they contain special characters. For example if the field name is Something Else!, then that can be specified like min(something-else-) or min([Something Else!]).

You can optionally specify the alias for the summarize expression using Summarize alias option. If nothing specified, summary will be used as alias. This alias will be helpful when you want to use merge transformation with results from different queries.