Keep dates and exposure totals honest

The sheet groups by the calendar date recorded in the mapped date keyword. It takes the YYYY-MM-DD part of an ISO-style date or timestamp. It does not convert to local time, infer your observatory's time zone or move post-midnight frames into the preceding observing night.

A midnight example

A frame recorded as 2026-09-12T23:58:00 belongs to September 12. A frame recorded as 2026-09-13T00:03:00 belongs to September 13. They form separate date groups even if you regard them as the same observing session. Check your capture software's date convention against your observing notes before correcting either value.

Missing or invalid dates block export for retained frames. Use a known acquisition date from your source records in the per-frame date input, or exclude the frame. No file modification time, current date, filename date or guessed session date is substituted.

Exposure is measured per frame

For 12 frames at 180 seconds and 4 at 300 seconds, the preview shows two rows. Total integration represented is (12 x 180) + (4 x 300) = 3,360 seconds. The CSV duration cells remain 180 and 300. Do not enter a group total into an individual frame's exposure input.

The default EXPTIME mapping assumes seconds, consistent with the FITS keyword examples in the FITS standard. Custom keyword names are not a unit conversion. If your instrument stores milliseconds, enter a verified per-frame seconds correction rather than mapping the raw milliseconds field as seconds.

Corrections remain inspectable

The audit keeps the original parsed header scalars and your per-frame date, filter and exposure overrides. It records the date policy and the exact frame IDs behind each group. It does not prove the correctness of your chosen corrections or validate clocks in the source instrument.

Continue to CSV import review or retained-frame selection.