LightCurveCollection.stitch() fails with TableMergeError when the collection has a mix of CDIPS and SPOC lightcurves

[orionlee]

Problem description

LightCurveCollection.stitch() fails with TableMergeError when the collection has a mix of CDIPS and SPOC lightcurves.

The culprit is that CDIPS lightcurve's quality column is type str while SPOC lightcurve's quality column is of type int.
(The issue would occur when there is a mix of CDIPS and most of the TESS HLSPs as well, as most of them have int type quality.)

The issue is similar to #954: the incompatibility here is different.

Example

import lightkurve as lk

# the tic has CDIPS in sector 11, SPOC in sector 64
sr = lk.search_lightcurve("TIC245486342", mission="TESS", sector=[11, 64])
sr = sr[np.isin(sr.author, ["SPOC", "CDIPS"])]
sr.download_all().stitch()

The error:

TableMergeError: The 'quality' columns have incompatible types: ['int32', 'str32']

Expected behavior

The incompatible columns due to different datatype, e.g., quality here, should be excluded from stitch. A warning should be issued (the behavior of issue #954 fix).

Environment

• lightkurve version (e.g. 1.0b6): 2.4.2