parse_to_columns should generate at least one row for a non-match (#7171

)
enso-org · Jun 30, 2023 · c866aa7 · c866aa7
1 parent 550d146
commit c866aa7
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 5 deletions.
diff --git a/distribution/lib/Standard/Table/0.0.0-dev/src/Internal/Split_Tokenize.enso b/distribution/lib/Standard/Table/0.0.0-dev/src/Internal/Split_Tokenize.enso
@@ -65,7 +65,7 @@ parse_to_columns table input_column_id pattern="." case_sensitivity=Case_Sensiti
     column = table.at input_column_id
 
     new_table = Value_Type.expect_text column <|
-        fan_out_to_rows_and_columns table input_column_id fun column_names on_problems=on_problems
+        fan_out_to_rows_and_columns table input_column_id fun column_names at_least_one_row=True on_problems=on_problems
     if parse_values then new_table.parse on_problems=on_problems else new_table
 
 ## PRIVATE

diff --git a/test/Table_Tests/src/In_Memory/Split_Tokenize_Spec.enso b/test/Table_Tests/src/In_Memory/Split_Tokenize_Spec.enso
@@ -306,20 +306,25 @@ spec =
             actual.should_equal expected
 
         Test.specify "input with no matches" <|
-            t = Table.from_rows ["foo", "bar", "baz"] [["x", "a", "y"]]
-            expected = Table.from_rows ["foo", "bar", "baz"] []
+            t = Table.from_rows ["foo", "bar", "baz"] [["x", "a", "y"], ["xx", "q12", "yy"], ["xxx", "34r", "yyyy"]]
+            expected = Table.from_rows ["foo", "bar", "baz"] [["x", Nothing, "y"], ["xx", 12, "yy"], ["xxx", 34, "yyyy"]]
             actual = t.parse_to_columns "bar" "\d+"
             actual.should_equal expected
 
+            t2 = Table.new [["amount", ["$1.23B", "$1.3M", "$2.32M $3.43B", "None"]]]
+            expected2 = Table.from_rows ["Raised", "Scale"] [[1.23, "B"], [1.3, "M"], [2.32, "M"], [3.43, "B"], [Nothing, Nothing]]
+            actual2 = t2.parse_to_columns "amount" "\$(?<Raised>\d+(?:.\d+)?)(?<Scale>B|M)"
+            actual2.should_equal expected2
+
         Test.specify "input with no matches, with regex groups" <|
             t = Table.from_rows ["foo", "bar", "baz"] [["x", "a", "y"]]
-            expected = Table.from_rows ["foo", "bar 1", "bar 2", "baz"] []
+            expected = Table.from_rows ["foo", "bar 1", "bar 2", "baz"] [["x", Nothing, Nothing, "y"]]
             actual = t.parse_to_columns "bar" "(\d)(\d)"
             actual.should_equal expected
 
         Test.specify "input with no matches, with named and unnamed regex groups" <|
             t = Table.from_rows ["foo", "bar", "baz"] [["x", "a", "y"]]
-            expected = Table.from_rows ["foo", "quux", "bar 1", "foo 1", "bar 2", "baz"] []
+            expected = Table.from_rows ["foo", "quux", "bar 1", "foo 1", "bar 2", "baz"] [["x", Nothing, Nothing, Nothing, Nothing, "y"]]
             actual = t.parse_to_columns "bar" "(?<quux>)(\d)(?<foo>\d)(\d)"
             actual.should_equal expected