Opposite of to_matrix_table_row_major()

Hail Query & hailctl
1

I have a MatrixTable which has only one row field (which is the row key “gene”) and one column field (which is the col key “s”). I have one float64 entry field so essentially I have a 2D MatrixTable. Is there a way to convert it into a Table so that I can export it to a pandas DataFrame?

2

It seems that make_table() is what I need but if I try to convert to a pandas dataframe it takes ages. My table should be about 512x10000 in size. It’s not huge.

3

@ag14774 I think you’re looking for localize_entries, which is an admittedly terrible name. You’ll probably want to do something like:

ht = mt.localize_entries('e', 'c')
column_metadata = ht.c.collect()[0][0]

ht = ht.transmute(**ht.e[0])
df = ht.to_pandas()