You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For example, I had a use case of viewing and processing iceberg snapshot ids (big int datatype). So the following query:
select snapshot_id from iceberg.iceberg_schema."iceberg_table$snapshots" order by committed_at
rounded off the actual values. [More details in the screenshots attached]
ISSUE: Bigint values are being rounded off in the UI, which leads to incorrect representation of data. For instance, a value of 1234567890123456623 may be displayed as 1234567890123456700, causing confusion and potential data misinterpretation.
This may be related to how the UI handles numeric types for display. A possible solution could be to update the UI rendering logic to handle bigint values correctly without rounding.
Load a dataset containing bigint values into Presto.
Query the dataset using a SQL statement (e.g., SELECT bigint_column FROM your_table).
Observe the results in the Presto UI.
Notice that the bigint values are rounded or displayed in scientific notation instead of their exact values.
A simpler reproduction step is SELECT 1234567890123456623, which will show as 1234567890123456500 in the UI. The CLI faithfully renders the original value.
After receiving the response from the server end, the JSON.parse() function is used to convert/parse the content into the regular object with names and values in proper data type. For bigint numbers, they are converted into Number data type in JavaScript and that's where the problem is. see the below example:
Need to find a workaround for this known issue in JavaScript. Currently, the JSON.parse() is in presto-js-client (here). We may need help from @alonsovb
Your Environment
Presto version used: 0.289-SNAPSHOT
Storage: File
Data source and connector used: Iceberg connector
Deployment: local
====================================================================================
Expected Behavior
Bigint values should be displayed accurately in the UI without rounding, reflecting their exact numeric representation.
====================================================================================
Current Behavior
For example, I had a use case of viewing and processing iceberg snapshot ids (big int datatype). So the following query:
select snapshot_id from iceberg.iceberg_schema."iceberg_table$snapshots" order by committed_at
rounded off the actual values. [More details in the screenshots attached]
ISSUE: Bigint values are being rounded off in the UI, which leads to incorrect representation of data. For instance, a value of 1234567890123456623 may be displayed as 1234567890123456700, causing confusion and potential data misinterpretation.
====================================================================================
Possible Solution
This may be related to how the UI handles numeric types for display. A possible solution could be to update the UI rendering logic to handle bigint values correctly without rounding.
====================================================================================
Steps to Reproduce
Load a dataset containing bigint values into Presto.
Query the dataset using a SQL statement (e.g., SELECT bigint_column FROM your_table).
Observe the results in the Presto UI.
Notice that the bigint values are rounded or displayed in scientific notation instead of their exact values.
====================================================================================
Screenshots
Presto-CLI console:
Presto UI:
The text was updated successfully, but these errors were encountered: