fix: sort dataframe key column in split

FelixAbrahamsson · FelixAbrahamsson · commit b3f045a3940b · 2021-02-23T10:30:46.000+01:00
diff --git a/datastream/tools/split_dataframes.py b/datastream/tools/split_dataframes.py
@@ -41,7 +41,7 @@ def split_dataframes(
             for split_name in proportions.keys()
         }
 
-    key_dataframe = pd.DataFrame({key_column: dataframe[key_column].unique()})
+    key_dataframe = pd.DataFrame({key_column: np.sort(dataframe[key_column].unique())})
 
     if frozen:
         if sum(map(len, split.values())) == 0:

Original file line number	Diff line number	Diff line change
`@@ -41,7 +41,7 @@ def split_dataframes(`
`41`	`41`	`for split_name in proportions.keys()`
`42`	`42`	`}`
`43`	`43`
`44`		`- key_dataframe = pd.DataFrame({key_column: dataframe[key_column].unique()})`
	`44`	`+ key_dataframe = pd.DataFrame({key_column: np.sort(dataframe[key_column].unique())})`
`45`	`45`
`46`	`46`	`if frozen:`
`47`	`47`	`if sum(map(len, split.values())) == 0:`