Hi Guys! Do you know what's wrong here? Custom created metric is not working

Hi Akmal, Thanks for sending this in, happy to help. The error message in your screenshot`"Missing aggregation, from='[attribute/...]' to='[]'"` indicates that your custom metric is missing an aggregation function when used with one or more attributes in the visualization. In GoodData, when you use a metric together with attributes (for example, in a bar chart grouped by categories), you must specify how the metric should be aggregated over those attributes (e.g., SUM, AVG, COUNT). This is a common requirement in GoodData MAQL and the UI: metrics must have an aggregation defined if they are to be sliced by attributes. If you do not specify an aggregation, GoodData cannot compute the metric for each attribute value, resulting in this error [Aggregation Functions]. How to fix: • Edit your custom metric and ensure you are using an aggregation function such as

SUM()

,

AVG()

,

COUNT()

, etc., around your fact or calculation. • For example, instead of

{fact/your_fact}

, use

SUM({fact/your_fact})

or another appropriate aggregation. If this doesn't help, kindly share a direct link to the report with the failing metric to allow us investigate and provide further assistance.

Hi @Yvonne Changamuka

Hi Akmal, Your metric is failing because of how you are using aggregation and conditional logic in MAQL. Here are the key issues to note; 1. DATETIME_DIFF Usage The

DATETIME_DIFF

function is not an aggregation function. When you use it in a metric that is sliced by attributes (like claim_id), you must specify the dataset context using the

BY

keyword. For example:

SELECT DATETIME_DIFF({label/cleared_date.day}, {label/posted_date.day}) BY {dataset/your_dataset}

Without the

BY

clause, GoodData does not know how to aggregate the difference for each row, which can cause errors, more information on this in DATETIME_DIFF documentation. 2. IF/CASE and Aggregation In MAQL, conditional statements like

IF

or

CASE

should be wrapped inside an aggregation function (such as

SUM

or

AVG

) when used in a metric that is sliced by attributes. This ensures the metric is computed at the correct dimensionality CASE documentation. 3. Nesting SELECT Statements You are nesting

SELECT

statements inside your

IF

condition, which is not the recommended MAQL pattern. Instead, you should use aggregations directly and avoid unnecessary nested

SELECT

statements. 4. COUNT and IFNULL The use of

IFNULL

is correct for handling nulls, but it should be applied to the result of an aggregation, not to a nested

SELECT

. How to Fix You should rewrite your metric to: • Use

DATETIME_DIFF

with the

BY

clause. • Wrap the

IF

statement inside an aggregation (e.g.,

SUM

or

AVG

).not nested selects IF syntax

SELECT IF ( <condition> , <then_expression> , <else_expression> )

• Avoid unnecessary nested

SELECT

statements. or equivalently:

SELECT CASE WHEN <condition> THEN <then_expression> ELSE <else_expression> END

You cannot write

SELECT IF <condition> THEN ... ELSE ... END

. Example Rewrite of how you can express what you intended (ratio of certain claims if dates are within 3 days)

SELECT

IF (

DATETIME_DIFF({label/cleared_date.day}, {label/posted_date.day}) <= 3,

(SELECT COUNT({label/claim_id})

WHERE {label/autopost_statuses_table.autopost_status} IN ("Posted", "Pre Posted"))

/

(SELECT COUNT({label/claim_id})),

NULL

)

Let me know if this helps. If not, kindly share a direct link to the report, feel free to share via dm. Thanks

Thank you so much @Yvonne Changamuka