1

Can someone provide the reason behind this feature?

For us, the column order is generally pretty sensible. We have lots of tables have grouped dimensions vs. metrics or IDs next to friendly names. This one-size-fits-all alphabetical approach is really unfortunate :crying_cat_face:

2

The reason was that in most cases the order that returns by default is random (based on the time column were added to the table) and alphabetical sort made more sense as it’s easier to reason about.

3

ah, it’s an interesting reason :grinning: thank you arik
and i mistakenly thought people should refrain from changing the schema so often :sweat_smile:

btw, i based my own fork off mozilla’s redash since they have the “Schema Viewer Drawer”. I’ve managed to restore the order using the “column description” field to store the ordinal position and sort on it.