From 709d6b41fc6ee5b81b602e60277ff4e5699db63e Mon Sep 17 00:00:00 2001
From: Sylvain Lesage <severo@rednegra.net>
Date: Fri, 5 Sep 2025 03:55:21 -0400
Subject: [PATCH] fix a bug in parquetQuery, when rowFormat is 'array' (#118)

It silently provided an empty array, instead of throwing an Error, or
providing the data in rowFormat="object".

Here, I (silently) force the rowFormat to "object".
---
 src/query.js       |  2 ++
 test/query.test.js | 24 ++++++++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/src/query.js b/src/query.js
index 816da66..b82fe7e 100644
--- a/src/query.js
+++ b/src/query.js
@@ -47,6 +47,7 @@ export async function parquetQuery(options) {
       // TODO: if expected > group size, start fetching next groups
       const groupData = await parquetReadObjects({
         ...options,
+        rowFormat: 'object',
         rowStart: groupStart,
         rowEnd: groupEnd,
         columns: relevantColumns,
@@ -71,6 +72,7 @@ export async function parquetQuery(options) {
     // read all rows, sort, and filter
     const results = await parquetReadObjects({
       ...options,
+      rowFormat: 'object',
       rowStart: undefined,
       rowEnd: undefined,
       columns: relevantColumns,
diff --git a/test/query.test.js b/test/query.test.js
index 2b3ae02..eeb478d 100644
--- a/test/query.test.js
+++ b/test/query.test.js
@@ -22,6 +22,18 @@ describe('parquetQuery', () => {
     ])
   })
 
+  it('returns rows in "array" format if asked', async () => {
+    const file = await asyncBufferFromFile('test/files/datapage_v2.snappy.parquet')
+    const rows = await parquetQuery({ file, rowFormat: 'array' })
+    expect(rows).toEqual([
+      [ 'abc', 1, 2, true, [1, 2, 3] ],
+      [ 'abc', 2, 3, true, undefined ],
+      [ 'abc', 3, 4, true, undefined ],
+      [ null, 4, 5, false, [1, 2, 3] ],
+      [ 'abc', 5, 2, true, [1, 2] ],
+    ])
+  })
+
   it('reads data with orderBy', async () => {
     const file = await asyncBufferFromFile('test/files/datapage_v2.snappy.parquet')
     const rows = await parquetQuery({ file, orderBy: 'c' })
@@ -63,6 +75,18 @@ describe('parquetQuery', () => {
     ])
   })
 
+  it('always returns rows in "object" format if filter is provided', async () => {
+    const file = await asyncBufferFromFile('test/files/datapage_v2.snappy.parquet')
+    const expected = [
+      { a: 'abc', b: 1, c: 2, d: true, e: [ 1, 2, 3 ] },
+      { a: 'abc', b: 5, c: 2, d: true, e: [ 1, 2 ] },
+    ]
+    const filter = { c: { $eq: 2 } }
+    expect(await parquetQuery({ file, filter, rowFormat: 'array' })).toEqual(expected)
+    expect(await parquetQuery({ file, filter, rowFormat: 'object' })).toEqual(expected)
+    expect(await parquetQuery({ file, filter })).toEqual(expected)
+  })
+
   it('reads data with filter and rowStart/rowEnd', async () => {
     const file = await asyncBufferFromFile('test/files/datapage_v2.snappy.parquet')
     const rows = await parquetQuery({ file, filter: { c: { $eq: 2 } }, rowStart: 1, rowEnd: 5 })