Fix bug in string column handling

jeromekelleher · Feb 7, 2024 · 2638cc3 · 2638cc3
1 parent e89e3fd
commit 2638cc3
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 5 deletions.
diff --git a/sgkit/io/vcf/vcf_converter.py b/sgkit/io/vcf/vcf_converter.py
@@ -138,10 +138,11 @@ def smallest_dtype(self):
             ret = "bool"
         else:
             assert self.vcf_type == "String"
-            if s.max_number == 0:
-                ret = "str"
-            else:
-                ret = "O"
+            ret = "str"
+            # if s.max_number == 0:
+            #     ret = "str"
+            # else:
+            #     ret = "O"
         # print("smallest dtype", self.name, self.vcf_type,":", ret)
         return ret
 

diff --git a/sgkit/tests/io/vcf/test_vcf_reader.py b/sgkit/tests/io/vcf/test_vcf_reader.py
@@ -1856,10 +1856,12 @@ def test_compare_vcf_to_zarr_convert(shared_datadir, tmp_path, vcf_name):
     # input for
     convert_vcf([vcf_path], zarr2_path)
     ds2 = load_dataset(zarr2_path)
-    vcf_to_zarr(vcf_path, zarr1_path, max_alt_alleles=ds2.variant_allele.shape[1] - 1)
+    vcf_to_zarr(vcf_path, zarr1_path, mixed_ploidy=True, max_alt_alleles=ds2.variant_allele.shape[1] - 1)
     ds1 = load_dataset(zarr1_path)
 
     # convert reads all variables by default.
     base_vars = list(ds1)
     ds2 = load_dataset(zarr2_path)
+    # print(ds1.call_genotype.values)
+    # print(ds2.call_genotype.values)
     xr.testing.assert_equal(ds1, ds2[base_vars])