Stroom Query Language

Query Format

Stroom Query Language (StroomQL) is a text based replacement for the existing Dashboard query builder and allows you to express the same queries in text form as well as providing additional functionality. It is currently used on the Query entity as the means of defining a query.

The following shows the supported syntax for a StroomQL query.

from <DATA_SOURCE>
where <FIELD> <CONDITION> <VALUE> [and|or|not]
[and|or|not]
[window] <TIME_FIELD> by <WINDOW_SIZE> [advance <ADVANCE_WINDOW_SIZE>]
[filter] <FIELD> <CONDITION> <VALUE> [and|or|not]
[and|or|not]
[eval...] <FIELD> = <EXPRESSION>
[having] <FIELD> <CONDITION> <VALUE> [and|or|not]
[group by] <FIELD>
[sort by] <FIELD> [desc|asc] // asc by default
[limit] <MAX_ROWS> 
select <FIELD> [as <COLUMN NAME>], ...
[show as] <VIS_NAME> (<VIS_CONTROL_ID_1> = <COLUMN_1>, <VIS_CONTROL_ID_2> = <COLUMN_2>)

Keywords

From

The first part of a StroomQL expression is the from clause that defines the single data source to query. All queries must include the from clause.

Select the data source to query, e.g.

from my_source

If the name of the data source contains white space then it must be quoted, e.g.

from "my source"

Where

Use where to construct query criteria, e.g.

where feed = "my feed"

Add boolean logic with and, or and not to build complex criteria, e.g.

where feed = "my feed"
or feed = "other feed"

Use brackets to group logical sub expressions, e.g.

where user = "bob"
and (feed = "my feed" or feed = "other feed")

Conditions

Supported conditions are:

• =
• !=
• >
• >=
• <
• <=
• is null
• is not null

And|Or|Not

Logical operators to add to where and filter clauses.

Bracket groups

You can force evaluation of items in a specific order using bracketed groups.

and X = 5 OR (name = foo and surname = bar)

Window

window <TIME_FIELD> by <WINDOW_SIZE> [advance <ADVANCE_WINDOW_SIZE>]

Windowing groups data by a specified window size applied to a time field. A window inserts additional rows for future periods so that rows for future periods contain count columns for previous periods.

Specify the field to window by and a duration. Durations are specified in simple terms e.g. 1d, 2w etc.

By default, a window will insert a count into the next period row. This is because by default we advance by the specified window size. If you wish to advance by a different duration you can specify the advance amount which will insert counts into multiple future rows.

Filter

Use filter to filter values that have not been indexed during search retrieval. This is used the same way as the where clause but applies to data after being retrieved from the index, e.g.

filter obscure_field = "some value"

Add boolean logic with and, or and not to build complex criteria as supported by the where clause. Use brackets to group logical sub expressions as supported by the where clause.

Eval

Use eval to assign the value returned from an Expression Function to a named variable, e.g.

eval my_count = count()

Here the result of the count() function is being stored in a variable called my_count. Functions can be nested and applied to variables, e.g.

eval new_name = concat(
  substring(name, 3, 5),
  substring(name, 8, 9))

Note that all fields in the data source selected using from will be available as variables by default.

Multiple eval statements can also be used to breakup complex function expressions and make it easier to comment out individual evaluations, e.g.

eval name_prefix = substring(name, 3, 5)
eval name_suffix = substring(name, 8, 9)
eval new_name = concat(
  name_prefix,
  name_suffix)

Variables can be reused, e.g.

eval name_prefix = substring(name, 3, 5)
eval new_name = substring(name, 8, 9)
eval new_name = concat(
  name_prefix,
  new_name)

In this example, the second assignment of new_name will override the value initially assigned to it. Note that that when reusing a variable name, the assignment can depend on the previous value assigned to that variable.

Add boolean logic with and, or and not to build complex criteria, e.g.

where feed = "my feed" or feed = "other feed"

Use brackets to group logical sub expressions, e.g.

where user = "bob" and (feed = "my feed" or feed = "other feed")

Having

A post aggregate filter that is applied at query time to return only rows that match the having conditions.

having count > 3

Group By

Use to group by columns, e.g.

group by feed

You can group across multiple columns, e.g.

group by feed, name

You can create nested groups, e.g.

group by feed
group by name

Sort By

Use to sort by columns, e.g.

sort by feed

You can sort across multiple columns, e.g.

sort by feed, name

You can change the sort direction, e.g.

sort by feed asc

Or

sort by feed desc

Limit

Limit the number of results, e.g.

limit 10

Select

The select keyword is used to define the fields that will be selected out of the data source (and any eval’d fields) for display in the table output.

select feed, name

You can optionally rename the fields so that they appear in the table with more human friendly names.

select feed as 'my feed column',
  name as 'my name column'

Show

The show keyword is used to tell StroomQL how to show the data resulting from the select. A Stroom visualisation can be specified and then passed column values from the select for the visualisation control properties.

show LineChart(x = EventTime, y = count)
show Doughnut(names = Feed, values = count)

For visualisations that contain spaces in their names it is necessary to use quotes, e.g.

show "My Visualisation" (x = EventTime, y = count)

Comments

Single line

StroomQL supports single line comments using //. For example:

from "index_view" // view
where EventTime > now() - 1227d
// and StreamId = 1210
select StreamId as "Stream Id", EventTime as "Event Time"

Multi line

Multiple lines can be commented by surrounding sections with /* and */. For example:

from "index_view" // view
where EventTime > now() - 1227d
/*
eval FirstName = lowerCase(substringBefore(UserId, '.'))
eval FirstName = any(FirstName)
*/
select StreamId as "Stream Id", EventTime as "Event Time"

Examples

The following are various example queries.

// add a where
from "index_view" // view
where EventTime > now() - 1227d
// and StreamId = 1210
eval UserId = any(upperCase(UserId))
eval FirstName = lowerCase(substringBefore(UserId, '.'))
eval FirstName = any(FirstName)
eval Sl = stringLength(FirstName)
eval count = count()
group by StreamId
sort by Sl desc
select Sl, StreamId as "Stream Id", EventId as "Event Id", EventTime as "Event Time", UserId as "User Id", FirstName, count
limit 10

from "index_view" // view
// add a where
where EventTime > now() - 1227d
// and StreamId = 1210
eval UserId = any(upperCase(UserId))
eval FirstName = lowerCase(substringBefore(UserId, '.'))
eval FirstName = any(FirstName)
eval Sl = stringLength(FirstName)
eval count = count()
group by StreamId
sort by Sl desc
select Sl, StreamId as "Stream Id", EventId as "Event Id", EventTime as "Event Time", UserId as "User Id", FirstName, count
limit 10

from "index_view" // view
// add a where
where EventTime > now() - 1227d
// and StreamId = 1210
eval UserId = any(upperCase(UserId))
eval FirstName = lowerCase(substringBefore(UserId, '.'))
eval FirstName = any(FirstName)
eval Sl = stringLength(FirstName)
// eval count = count()
// group by StreamId
// sort by Sl desc
select StreamId as "Stream Id", EventId as "Event Id"
// limit 10

from "index_view" // view
// add a where
where EventTime > now() - 1227d
// and StreamId = 1210
eval UserId = any(upperCase(UserId))
eval FirstName = lowerCase(substringBefore(UserId, '.'))
eval FirstName = any(FirstName)
eval Sl = stringLength(FirstName)
eval count = count()
group by StreamId
sort by Sl desc
select Sl, StreamId as "Stream Id", EventId as "Event Id", EventTime as "Event Time", UserId as "User Id", FirstName, count
limit 10