Search data commands reference

This reference identifies the commands you can use to modify search data query generation at cluster-level. These commands must be enabled by ThoughtSpot Support.

Commands

Category Name Description

Formulas

Enable / disable case-sensitive formulas

The default setting is TRUE. When TRUE, the string case for formula definitions is saved as lower case. When FALSE, the case entered in the formula definition is respected. The ability to respect the case is important when leveraging passthrough functions. This is because functions defined in the underlying data platforms are often case-sensitive.

Example when true Entered formula = if(city_name = “New York”) then ‘YES’ else ‘NO’ Saved formula = if(city_name = “new york”) then ‘yes’ else ‘no’

Example when false Entered formula = if(city_name = “New York”) then ‘YES’ else ‘NO’ Saved formula = if(city_name = “New York”) then ‘YES’ else ‘NO’

Query generation

Measures in case statements considered zero if null

The default setting is TRUE. When TRUE, all measures are wrapped in a case statement to consider NULL as ZERO. When FALSE, the underlying query is simplified.

Example when true

SELECT CASE
WHEN sum("ta_1"."Sales_Dollar_Amount") IS NOT NULL THEN sum("ta_1"."Sales_Dollar_Amount")
ELSE 0
END "ca_1"
FROM "HO_RETAIL"."PUBLIC"."HO_RETAIL_SALES_FACT" "ta_1"
ORDER BY "ca_1" ASC NULLS LAST

Example when false

SELECT sum("ta_1"."Sales_Dollar_Amount") "ca_1"
FROM "HO_RETAIL"."PUBLIC"."HO_RETAIL_SALES_FACT" "ta_1"

Query generation

Non-shared filters in chasm trap

The default setting is FALSE. This setting affects how non-shared filters are applied in a chasm trap model. That is, the filter is only applied to one side of the chasm trap. Assume a filter is applied to the LEFT HAND side of the chasm trap.

When TRUE, the query is a FULL OUTER JOIN. When FALSE, the resulting query is a LEFT OUTER JOIN. This is normally a much more performance query.

Where multiple filters are applied from both sides of the chasm trap, then ThoughtSpot will intelligently determine the join type.

This article provides further details when this setting is FALSE.

Row-level security

Apply RLS rules on joined tables

The default setting is TRUE. Assume a model where fact1 joins to dimension1 and an RLS rule is applied to fact.

When TRUE, if a column is selected from dimension1 then a join to fact1 is included. That is, the RLS rule is applied. In this case, only the values from dimension1 which meet the rule on fact1 are returned. When FALSE, if a column is selected from dimension1, then no RLS rule is applied. That is, all the values are returned as the results for the query.

Refer to this video for further information regarding row level security.

Row-level security

Accept parent groups as ts_groups() values in RLS

When FALSE, only immediate groups are values of ts_groups(). When TRUE, parent groups are also included.

Pivot table

Calculate pivot table summaries as separate SQL queries

The default setting is TRUE. When FALSE, the pivot table calculations are calculated in the front end. These results are limited to a SUM of the numbers displayed in the pivot table. This is also limited to 100k rows of data.

When TRUE, the summaries are calculated as separate SQL queries. This supports the ability to define different aggregation rules. For example, Average, Min, Max.

Column-level security

Implement column-level security

The default setting is FALSE. To implement column-level security, this setting should be set to TRUE. This results in a stricter object security model. In short, all data objects must be shared with consumers rather than just Worksheets. Refer to this article and this video.

Column-level security — UI controls

In column-level security, hide underlying tables

The default setting is FALSE. This setting is used in conjunction with the setting for column-level security. That is, with column-level security, the underlying tables must be shared with consumers. When TRUE, this ensures that the underlying tables are not available as data sources from the search data source selection. That is, consumers are limited to selecting Worksheets or Views.


Was this page helpful?