Hi @szymon_dybczak, thank you for your response. 

If I update the table property to include all 42 columns (new 10 + old 32 cols) in delta.dataSkippingStatsColumns,

and then run ANALYZE TABLE only for the new 10 columns (col33–col42):

My Questions:

• Will this keep the existing statistics for columns 1–32 unchanged and only compute statistics for columns 33–42?
• Are the statistics for both the old (1–32) and new (33–42) columns maintained together in the same centralized metadata location, or are they managed separately?
• Does this approach ensure that all 42 columns will have statistics available for data skipping without any redundant recomputation? (i.e., existing stats stay as-is, and only new stats are added)
• After setting the dataSkipping property for all 42 columns, will future data loads automatically generate and maintain statistics for all 42 columns going forward?

Appreciate any insights and clarifications!