fix: Expand brackets in async glob expansion (#17630)

crates/polars-io/src/cloud/glob.rs

@@ -19,8 +19,7 @@ fn extract_prefix_expansion(url: &str) -> PolarsResult<(String, Option<String>)>
     let mut expansion = String::new();
     let mut last_split_was_wildcard = false;
     for split in splits {
-        let has_star = split.contains('*');
-        if expansion.is_empty() && !has_star {
+        if expansion.is_empty() && memchr::memchr2(b'*', b'[', split.as_bytes()).is_none() {
             // We are still gathering splits in the prefix.
             if !prefix.is_empty() {
                 prefix.push(DELIMITER);
@@ -44,7 +43,7 @@ fn extract_prefix_expansion(url: &str) -> PolarsResult<(String, Option<String>)>
             expansion.push(DELIMITER);
         }
         // Handle '.' inside a split.
-        if split.contains('.') || split.contains('*') {
+        if memchr::memchr2(b'.', b'*', split.as_bytes()).is_some() {
             let processed = split.replace('.', "\\.");
             expansion.push_str(&processed.replace('*', "([^/]*)"));
             continue;

crates/polars-io/src/utils/path.rs

@@ -315,7 +315,7 @@ pub fn expand_paths_hive(
                 if path.extension() != ext {
                     polars_bail!(
                         InvalidOperation: r#"directory contained paths with different file extensions: \
-                        first path: {}, second path: {}. Please use a glob pattern to explicitly specify
+                        first path: {}, second path: {}. Please use a glob pattern to explicitly specify \
                         which files to read (e.g. "dir/**/*", "dir/**/*.parquet")"#,
                         out_paths[i - 1].to_str().unwrap(), path.to_str().unwrap()
                     );

py-polars/tests/unit/io/test_scan.py

@@ -655,3 +655,13 @@ def test_scan_include_file_name(
 
     # Test codepaths that materialize empty DataFrames
     assert_frame_equal(lf.head(0).collect(streaming=streaming), df.head(0))
+
+
+@pytest.mark.write_disk()
+def test_async_path_expansion_bracket_17629(tmp_path: Path) -> None:
+    path = tmp_path / "data.parquet"
+
+    df = pl.DataFrame({"x": 1})
+    df.write_parquet(path)
+
+    assert_frame_equal(pl.scan_parquet(tmp_path / "[d]ata.parquet").collect(), df)