From 6758bc2bd7c5149f915034e130e8a004b5bf5fd7 Mon Sep 17 00:00:00 2001
From: Daniel McDonald <danielmcdonald@ucsd.edu>
Date: Tue, 19 Sep 2023 15:59:34 -0700
Subject: [PATCH 1/2] Fixes #149

---
 q2_feature_table/_filter.py    | 5 +----
 q2_feature_table/_subsample.py | 2 +-
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/q2_feature_table/_filter.py b/q2_feature_table/_filter.py
index 8253587..b707902 100644
--- a/q2_feature_table/_filter.py
+++ b/q2_feature_table/_filter.py
@@ -57,10 +57,7 @@ def _filter_table(table, min_frequency, max_frequency, min_nonzero,
     # filter on the opposite axis to remove any entities that now have a
     # frequency of zero
     if filter_opposite_axis:
-        filter_fn2 = _get_biom_filter_function(
-            ids_to_keep=table.ids(axis=_other_axis_map[axis]), min_frequency=0,
-            max_frequency=None, min_nonzero=1, max_nonzero=None)
-        table.filter(filter_fn2, axis=_other_axis_map[axis], inplace=True)
+        table.remove_empty(axis=_other_axis_map[axis], inplace=True)
 
 
 def filter_samples(table: biom.Table, min_frequency: int = 0,
diff --git a/q2_feature_table/_subsample.py b/q2_feature_table/_subsample.py
index 7a29293..a2d353d 100644
--- a/q2_feature_table/_subsample.py
+++ b/q2_feature_table/_subsample.py
@@ -25,7 +25,7 @@ def subsample(table: biom.Table, subsampling_depth: int,
 
     # the inverted axis is always observation due to the above transpose
     invaxis = 'observation'
-    table.filter(lambda v, i, m: v.sum() > 0, axis=invaxis)
+    table = table.remove_empty(axis=invaxis, inplace=False)
 
     if axis == 'feature':
         # reverse the transpose necessary due to biocore/biom-format#759

From 7a9a89f84225e8f44b1669d0015cb15e6aefded2 Mon Sep 17 00:00:00 2001
From: Daniel McDonald <danielmcdonald@ucsd.edu>
Date: Tue, 19 Sep 2023 16:15:58 -0700
Subject: [PATCH 2/2] MAINT: use Table.partition

---
 q2_feature_table/_split.py | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

diff --git a/q2_feature_table/_split.py b/q2_feature_table/_split.py
index a7d846d..ebf0775 100644
--- a/q2_feature_table/_split.py
+++ b/q2_feature_table/_split.py
@@ -14,12 +14,14 @@ def split(table: biom.Table,
           metadata: qiime2.CategoricalMetadataColumn,
           filter_empty_features: bool = True) -> biom.Table:
     metadata_df = metadata.drop_missing_values().to_dataframe()
+    lookup = metadata_df[metadata.name].to_dict()
 
-    indices = metadata_df.reset_index(
-        ).groupby(metadata.name)[metadata_df.index.name].apply(list).to_dict()
+    def partition_f(i, m):
+        return lookup.get(i)
 
+    unique_grps = sorted(set(lookup.values()))
     try:
-        qiime2.sdk.util.validate_result_collection_keys(*indices.keys())
+        qiime2.sdk.util.validate_result_collection_keys(*unique_grps)
     except KeyError as e:
         raise KeyError(
             "One or more invalid metadata column values identified during "
@@ -28,9 +30,11 @@ def split(table: biom.Table,
             f"table. The original error message is as follows: {str(e)}")
 
     result = {}
-    for group, sample_ids in indices.items():
-        t = table.filter(sample_ids, axis='sample', inplace=False)
+    for group, tab in table.partition(partition_f):
+        if group is None:
+            continue
+
         if filter_empty_features:
-            t.remove_empty(axis='observation', inplace=True)
-        result[group] = t
+            tab.remove_empty(axis='observation', inplace=True)
+        result[group] = tab
     return result