problem about filter_rank_genes_groups

Hi all,
I run filter_rank_genes_groups follow this https://scanpy.readthedocs.io/en/stable/generated/scanpy.tl.filter_rank_genes_groups.html

import scanpy as sc
adata = sc.datasets.pbmc68k_reduced()
sc.tl.rank_genes_groups(adata, 'bulk_labels', method='wilcoxon')
sc.tl.filter_rank_genes_groups(adata, min_fold_change=3)
# visualize results
sc.pl.rank_genes_groups(adata, key='rank_genes_groups_filtered')
# visualize results using dotplot
sc.pl.rank_genes_groups_dotplot(adata, key='rank_genes_groups_filtered')

However, my gene names become NAN.
What should I do to solve this problem?

adata.uns['rank_genes_groups_filtered']

{'params': {'groupby': 'leiden_0.4', 'reference': 'rest', 'method': 'wilcoxon', 'use_raw': False, 'layer': None, 'corr_method': 'benjamini-hochberg'}, 'pts': 0 1 2 3 4 5
AL627309.1 1.0 1.0 1.0 1.0 1.0 1.0
RP11-206L10.2 1.0 1.0 1.0 1.0 1.0 1.0
LINC00115 1.0 1.0 1.0 1.0 1.0 1.0
NOC2L 1.0 1.0 1.0 1.0 1.0 1.0
KLHL17 1.0 1.0 1.0 1.0 1.0 1.0
... ... ... ... ... ... ...
AC145212.1 1.0 1.0 1.0 1.0 1.0 1.0
AL592183.1 1.0 1.0 1.0 1.0 1.0 1.0
AL354822.1 1.0 1.0 1.0 1.0 1.0 1.0
PNRC2-1 1.0 1.0 1.0 1.0 1.0 1.0
SRSF10-1 1.0 1.0 1.0 1.0 1.0 1.0

[12381 rows x 6 columns], 'pts_rest': 0 1 2 3 4 5
AL627309.1 1.0 1.0 1.0 1.0 1.0 1.0
RP11-206L10.2 1.0 1.0 1.0 1.0 1.0 1.0
LINC00115 1.0 1.0 1.0 1.0 1.0 1.0
NOC2L 1.0 1.0 1.0 1.0 1.0 1.0
KLHL17 1.0 1.0 1.0 1.0 1.0 1.0
... ... ... ... ... ... ...
AC145212.1 1.0 1.0 1.0 1.0 1.0 1.0
AL592183.1 1.0 1.0 1.0 1.0 1.0 1.0
AL354822.1 1.0 1.0 1.0 1.0 1.0 1.0
PNRC2-1 1.0 1.0 1.0 1.0 1.0 1.0
SRSF10-1 1.0 1.0 1.0 1.0 1.0 1.0

[12381 rows x 6 columns], 'names': rec.array([(nan, nan, nan, nan, nan, nan), (nan, nan, nan, nan, nan, nan),
(nan, nan, nan, nan, nan, nan), ...,
(nan, nan, nan, nan, nan, nan), (nan, nan, nan, nan, nan, nan),
(nan, nan, nan, nan, nan, nan)],
dtype=[('0', 'O'), ('1', 'O'), ('2', 'O'), ('3', 'O'), ('4', 'O'), ('5', 'O')]), 'scores': rec.array([( 31.145878, 33.94006 , 30.770067, 29.246338, 16.453491, 5.9842377),
( 27.875402, 32.776535, 26.378975, 28.712181, 16.096996, 5.9842377),
( 27.05427 , 32.579735, 25.802261, 27.495918, 15.874588, 5.983843 ),
...,
(-26.778828, -24.484423, -14.373905, -17.513113, -13.714344, -5.740007 ),
(-27.320658, -27.529768, -14.619217, -19.539616, -13.793954, -5.828585 ),
(-27.61369 , -31.664293, -16.766815, -20.83902 , -14.882286, -5.9672713)],
dtype=[('0', '<f4'), ('1', '<f4'), ('2', '<f4'), ('3', '<f4'), ('4', '<f4'), ('5', '<f4')]), 'pvals': rec.array([(5.76757182e-213, 1.70978363e-252, 6.59163276e-208, 4.99861075e-188, 7.91707478e-61, 2.17405930e-09),
(5.30272154e-171, 1.27177903e-235, 2.38824940e-153, 2.68829460e-181, 2.67790267e-58, 2.17405930e-09),
(3.40236053e-161, 7.94325574e-233, 8.36374334e-147, 1.96443112e-166, 9.50323353e-57, 2.17933501e-09),
...,
(5.70224910e-158, 2.16463296e-132, 7.54533758e-047, 1.13797729e-068, 8.33159464e-43, 9.46726154e-09),
(2.41092484e-164, 7.73129397e-167, 2.11823139e-048, 5.05610163e-085, 2.77143736e-43, 5.58993361e-09),
(7.62160014e-168, 4.82099981e-220, 4.26779582e-063, 1.91722780e-096, 4.29568606e-50, 2.41253723e-09)],
dtype=[('0', '<f8'), ('1', '<f8'), ('2', '<f8'), ('3', '<f8'), ('4', '<f8'), ('5', '<f8')]), 'pvals_adj': rec.array([(7.14083067e-209, 2.11688312e-248, 8.16110052e-204, 6.18877997e-184, 9.80213028e-57, 2.99072523e-06),
(3.28264977e-167, 7.87294806e-232, 1.47844579e-149, 1.66418877e-177, 1.65775565e-54, 2.99072523e-06),
(8.42492515e-158, 3.27818164e-229, 3.45171688e-143, 8.10720722e-163, 3.92198448e-53, 2.99072523e-06),
...,
(1.17665910e-154, 1.11668003e-129, 5.18993470e-044, 8.28782167e-066, 8.59612277e-40, 6.51189806e-06),
(7.46241510e-161, 5.63065592e-164, 1.63911392e-045, 4.17330628e-082, 3.43131660e-40, 4.94349772e-06),
(3.14543438e-164, 9.94813310e-217, 4.80359819e-060, 1.82593826e-093, 8.86414818e-47, 2.99072523e-06)],
dtype=[('0', '<f8'), ('1', '<f8'), ('2', '<f8'), ('3', '<f8'), ('4', '<f8'), ('5', '<f8')]), 'logfoldchanges': rec.array([(nan, nan, nan, nan, nan, nan), (nan, nan, nan, nan, nan, nan),
(nan, nan, nan, nan, nan, nan), ...,
(nan, nan, nan, nan, nan, nan), (nan, nan, nan, nan, nan, nan),
(nan, nan, nan, nan, nan, nan)],
dtype=[('0', '<f4'), ('1', '<f4'), ('2', '<f4'), ('3', '<f4'), ('4', '<f4'), ('5', '<f4')])}

Start by reproducing the example from the linked filter_rank_genes_groups documentation with sc.datasets.pbmc68k_reduced(), then inspect adata.uns['rank_genes_groups_filtered']. Compare the filtered names and logfoldchanges with the unfiltered rank_genes_groups result. Done means identifying why the filtered output contains NaN values and confirming the corrected output or documenting the expected behavior.