From e0c107e83ab05ead1341b36868ef0f35503b989d Mon Sep 17 00:00:00 2001
From: Noah Treuhaft <noah.treuhaft@gmail.com>
Date: Mon, 24 Aug 2026 13:23:38 -0400
Subject: [PATCH] sio/parquetio: fix projection pushdown lists and maps

Projection pushdown fails for lists and maps because columnIndexes
returns nothing for columns with nested logical types (LIST and MAP).
It does that because it looks up columns via
parquet/schema.Schema.ColumnIndexByName, which exposes the structuring
fields not present in the data (i.e., list and element for LIST; map,
key, and value for MAP).  Fix this by finding column indexes via
pqarrow.SchemaManifest, which does not include those fields.
---
 sio/parquetio/reader.go              | 31 ++++++++++++++++++++++------
 sio/parquetio/ztests/issue-7220.yaml | 14 +++++++++++++
 2 files changed, 39 insertions(+), 6 deletions(-)
 create mode 100644 sio/parquetio/ztests/issue-7220.yaml

diff --git a/sio/parquetio/reader.go b/sio/parquetio/reader.go
index 09c0ea7d99..24eb17cc6c 100644
--- a/sio/parquetio/reader.go
+++ b/sio/parquetio/reader.go
@@ -15,7 +15,6 @@ import (
 	"github.com/apache/arrow-go/v18/parquet"
 	"github.com/apache/arrow-go/v18/parquet/file"
 	"github.com/apache/arrow-go/v18/parquet/pqarrow"
-	"github.com/apache/arrow-go/v18/parquet/schema"
 	"github.com/brimdata/super"
 	"github.com/brimdata/super/pkg/byteconv"
 	"github.com/brimdata/super/pkg/field"
@@ -84,7 +83,7 @@ func NewReader(ctx context.Context, sctx *super.Context, r io.Reader, p sbuf.Pus
 				// Trim trailing "max" or "min".
 				paths[i] = p[:len(p)-1]
 			}
-			colIndexes := columnIndexes(pr.MetaData().Schema, paths)
+			colIndexes := columnIndexes(schemaManifest, paths)
 			// Remove duplicates created above by trimming "max" and "min".
 			metadataColIndexes = slices.Compact(colIndexes)
 			for range concurrentReaders {
@@ -104,7 +103,7 @@ func NewReader(ctx context.Context, sctx *super.Context, r io.Reader, p sbuf.Pus
 		ctx:                ctx,
 		sctx:               sctx,
 		fr:                 fr,
-		colIndexes:         columnIndexes(prmd.Schema, fields),
+		colIndexes:         columnIndexes(schemaManifest, fields),
 		colIndexToField:    schemaManifest.ColIndexToField,
 		metadataColIndexes: metadataColIndexes,
 		metadataFilters:    metadataFilters,
@@ -114,16 +113,36 @@ func NewReader(ctx context.Context, sctx *super.Context, r io.Reader, p sbuf.Pus
 	}, nil
 }
 
-func columnIndexes(schema *schema.Schema, fields []field.Path) []int {
+func columnIndexes(sm *pqarrow.SchemaManifest, fields []field.Path) []int {
 	var indexes []int
 	for _, f := range fields {
-		if i := schema.ColumnIndexByName(f.String()); i >= 0 {
-			indexes = append(indexes, i)
+		indexes = appendColumnIndexesForField(indexes, sm.Fields, f)
+	}
+	return indexes
+}
+
+func appendColumnIndexesForField(indexes []int, sfs []pqarrow.SchemaField, f field.Path) []int {
+	for _, sf := range sfs {
+		if sf.Field.Name == f[0] {
+			if len(f) == 1 {
+				return appendColumnIndexes(indexes, &sf)
+			}
+			return appendColumnIndexesForField(indexes, sf.Children, f[1:])
 		}
 	}
 	return indexes
 }
 
+func appendColumnIndexes(indexes []int, sf *pqarrow.SchemaField) []int {
+	if sf.IsLeaf() {
+		return append(indexes, sf.ColIndex)
+	}
+	for _, sf := range sf.Children {
+		indexes = appendColumnIndexes(indexes, &sf)
+	}
+	return indexes
+}
+
 func (r *Reader) Pull(done bool) (vector.Any, error) {
 	return r.ConcurrentPull(done, 0)
 }
diff --git a/sio/parquetio/ztests/issue-7220.yaml b/sio/parquetio/ztests/issue-7220.yaml
new file mode 100644
index 0000000000..f52075a1e2
--- /dev/null
+++ b/sio/parquetio/ztests/issue-7220.yaml
@@ -0,0 +1,14 @@
+# Verify that projection pushdown works with nested logical types.
+
+script: |
+  super -f parquet -o a.parquet -
+  super -s -c 'from a.parquet | values {a,b,c}'
+
+inputs:
+  - name: stdin
+    data: &stdin |
+      {a:1,b:[2],c:map{"cc":3}}
+
+outputs:
+  - name: stdout
+    data: *stdin