fix: Capture groups should be ignored in replace when literal=True (#…

…19413)

crates/polars-plan/src/dsl/function_expr/strings.rs

@@ -844,20 +844,26 @@ fn replace_n<'a>(
                 "replacement value length ({}) does not match string column length ({})",
                 len_val, ca.len(),
             );
-            let literal = literal || is_literal_pat(&pat);
+            let lit = is_literal_pat(&pat);
+            let literal_pat = literal || lit;
 
-            if literal {
+            if literal_pat {
                 pat = escape(&pat)
             }
 
             let reg = Regex::new(&pat)?;
-            let lit = pat.chars().all(|c| !c.is_ascii_punctuation());
 
             let f = |s: &'a str, val: &'a str| {
                 if lit && (s.len() <= 32) {
                     Cow::Owned(s.replacen(&pat, val, 1))
                 } else {
-                    reg.replace(s, val)
+                    // According to the docs for replace
+                    // when literal = True then capture groups are ignored.
+                    if literal {
+                        reg.replace(s, NoExpand(val))
+                    } else {
+                        reg.replace(s, val)
+                    }
                 }
             };
             Ok(iter_and_replace(ca, val, f))

py-polars/tests/unit/operations/namespaces/string/test_string.py

@@ -1006,7 +1006,7 @@ def test_replace_all() -> None:
     )
 
 
-def test_replace_literal_no_caputures() -> None:
+def test_replace_all_literal_no_caputures() -> None:
     # When using literal = True, capture groups should be disabled
 
     # Single row code path in Rust
@@ -1034,6 +1034,38 @@ def test_replace_literal_no_caputures() -> None:
     assert df2.get_column("text2")[1] == "I lost $2 yesterday."
 
 
+def test_replace_literal_no_caputures() -> None:
+    # When using literal = True, capture groups should be disabled
+
+    # Single row code path in Rust
+    df = pl.DataFrame({"text": ["I found <amt> yesterday."], "amt": ["$1"]})
+    df = df.with_columns(
+        pl.col("text").str.replace("<amt>", pl.col("amt"), literal=True).alias("text2")
+    )
+    assert df.get_column("text2")[0] == "I found $1 yesterday."
+
+    # Multi-row code path in Rust
+    # A string shorter than 32 chars,
+    # and one longer than 32 chars to test both sub-paths
+    df2 = pl.DataFrame(
+        {
+            "text": [
+                "I found <amt> yesterday.",
+                "I lost <amt> yesterday and this string is longer than 32 characters.",
+            ],
+            "amt": ["$1", "$2"],
+        }
+    )
+    df2 = df2.with_columns(
+        pl.col("text").str.replace("<amt>", pl.col("amt"), literal=True).alias("text2")
+    )
+    assert df2.get_column("text2")[0] == "I found $1 yesterday."
+    assert (
+        df2.get_column("text2")[1]
+        == "I lost $2 yesterday and this string is longer than 32 characters."
+    )
+
+
 def test_replace_expressions() -> None:
     df = pl.DataFrame({"foo": ["123 bla 45 asd", "xyz 678 910t"], "value": ["A", "B"]})
     out = df.select([pl.col("foo").str.replace(pl.col("foo").first(), pl.col("value"))])