Welcome! I've added you as a Contributor in JIRA and assigned IMPALA-12918 to you.
The approach sounds reasonable. I assume for the moment this will not change how we handle these stats when reading them, just the validation when writing them with ALTER TABLE. On Tue, Nov 25, 2025 at 8:54 PM KunalSiyag <[email protected]> wrote: > Hello Impala Developers, > > My name is Kunal Siyag, and I am a new contributor to the Apache Impala > project. > > I have registered an Apache JIRA account with the username *dazz*. > > I would like to work on the JIRA ticket *[IMPALA-12918] - Do not allow > non-numeric values in Hive table stats during an alter table*. I see it > is currently unassigned and marked as a ramp-up/newbie issue, and I > intend to start working on it shortly. > > Since I am a new contributor and do not have committer permissions, > could one of the committers please: > > *Grant my user access to assign JIRAs* (as mentioned in the > contribution guide). > > *Assign IMPALA-12918* to my JIRA account, *dazz*. > > I have also included a brief outline of my intended approach below for > discussion (as recommended for all but the very smallest items): > > Intended Approach > The core issue is that ALTER TABLE ... SET TBLPROPERTIES allows > non-numeric strings for properties like numRows, totalSize, and > rawDataSize, which should only be numeric. > > My plan is to implement the validation within the Catalog component, > specifically where the table properties are processed during an ALTER > TABLE statement: > > *Identify Target Properties:* Create a static list of the table > properties that require numeric validation (numRows, totalSize, > rawDataSize). > > *Add Validation:* During the parsing/processing of the SET > TBLPROPERTIES clause, check if the value provided for any of the target > properties is a valid number. > > *Throw Error:* If a property value is non-numeric (including empty > string), throw an informative exception to the user, preventing the > alter table operation from completing. > > *Testing:* Add new test cases to cover the successful numeric inputs > and the new validation failures for non-numeric/empty inputs. I will > also check existing Beeline/Hive behavior as noted in the JIRA > description. > > Please let me know if this approach aligns with the project's goals or > if there's an existing utility or location where this validation should > be placed. > > Thank you for your help in getting me started! > > Best regards, > > Kunal Siyag > > dazz > > >
