From de03425587cfd6043c33b8ba93876c240b528a66 Mon Sep 17 00:00:00 2001
From: Kenny Daniel <platypii@gmail.com>
Date: Thu, 17 Apr 2025 01:09:43 -0700
Subject: [PATCH] Don't write file_path, duckdb doesn't like it

---
 src/parquet-writer.js     |  3 +--
 test/example.js           |  9 +--------
 test/metadata.test.js     |  2 +-
 test/write.buffer.test.js | 14 +++++++-------
 4 files changed, 10 insertions(+), 18 deletions(-)

diff --git a/src/parquet-writer.js b/src/parquet-writer.js
index ed198db..3fbede6 100644
--- a/src/parquet-writer.js
+++ b/src/parquet-writer.js
@@ -48,7 +48,7 @@ ParquetWriter.prototype.write = function({ columnData, rowGroupSize = 100000 })
 
     // write columns
     for (let j = 0; j < columnData.length; j++) {
-      const { name, data } = columnData[j]
+      const { data } = columnData[j]
       const schemaPath = [this.schema[0], this.schema[j + 1]]
       const groupData = data.slice(groupStartIndex, groupStartIndex + groupSize)
       const file_offset = BigInt(this.writer.offset)
@@ -56,7 +56,6 @@ ParquetWriter.prototype.write = function({ columnData, rowGroupSize = 100000 })
 
       // save column chunk metadata
       columns.push({
-        file_path: name,
         file_offset,
         meta_data,
       })
diff --git a/test/example.js b/test/example.js
index 4433488..866aadc 100644
--- a/test/example.js
+++ b/test/example.js
@@ -31,7 +31,6 @@ export const exampleMetadata = {
   row_groups: [{
     columns: [
       {
-        file_path: 'bool',
         file_offset: 4n,
         meta_data: {
           type: 'BOOLEAN',
@@ -50,7 +49,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'int',
         file_offset: 28n,
         meta_data: {
           type: 'INT32',
@@ -69,7 +67,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'bigint',
         file_offset: 67n,
         meta_data: {
           type: 'INT64',
@@ -88,7 +85,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'float',
         file_offset: 110n,
         meta_data: {
           type: 'FLOAT',
@@ -107,7 +103,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'double',
         file_offset: 149n,
         meta_data: {
           type: 'DOUBLE',
@@ -126,7 +121,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'string',
         file_offset: 200n,
         meta_data: {
           type: 'BYTE_ARRAY',
@@ -145,7 +139,6 @@ export const exampleMetadata = {
         },
       },
       {
-        file_path: 'nullable',
         file_offset: 242n,
         meta_data: {
           type: 'BOOLEAN',
@@ -167,5 +160,5 @@ export const exampleMetadata = {
     total_byte_size: 264n,
     num_rows: 4n,
   }],
-  metadata_length: 497,
+  metadata_length: 445,
 }
diff --git a/test/metadata.test.js b/test/metadata.test.js
index 527b493..9f4d009 100644
--- a/test/metadata.test.js
+++ b/test/metadata.test.js
@@ -24,7 +24,7 @@ describe('writeMetadata', () => {
         { key: 'key1', value: 'value1' },
         { key: 'key2', value: 'value2' },
       ],
-      metadata_length: 529,
+      metadata_length: 477,
     }
     writeMetadata(writer, withKvMetadata)
 
diff --git a/test/write.buffer.test.js b/test/write.buffer.test.js
index c7b4864..c2f600e 100644
--- a/test/write.buffer.test.js
+++ b/test/write.buffer.test.js
@@ -39,9 +39,9 @@ describe('parquetWriteBuffer', () => {
     bool[500] = true
     bool[9999] = false
     const file = parquetWriteBuffer({ columnData: [{ name: 'bool', data: bool }] })
-    expect(file.byteLength).toBe(160)
+    expect(file.byteLength).toBe(154)
     const metadata = parquetMetadata(file)
-    expect(metadata.metadata_length).toBe(98)
+    expect(metadata.metadata_length).toBe(92)
     const result = await parquetReadObjects({ file })
     expect(result.length).toBe(10000)
     expect(result[0]).toEqual({ bool: null })
@@ -55,14 +55,14 @@ describe('parquetWriteBuffer', () => {
   it('efficiently serializes long string', () => {
     const str = 'a'.repeat(10000)
     const file = parquetWriteBuffer({ columnData: [{ name: 'string', data: [str] }] })
-    expect(file.byteLength).toBe(646)
+    expect(file.byteLength).toBe(638)
   })
 
   it('less efficiently serializes string without compression', () => {
     const str = 'a'.repeat(10000)
     const columnData = [{ name: 'string', data: [str] }]
     const file = parquetWriteBuffer({ columnData, compressed: false })
-    expect(file.byteLength).toBe(10176)
+    expect(file.byteLength).toBe(10168)
   })
 
   it('efficiently serializes column with few distinct values', async () => {
@@ -70,7 +70,7 @@ describe('parquetWriteBuffer', () => {
       .fill('aaaa', 0, 50000)
       .fill('bbbb', 50000, 100000)
     const file = parquetWriteBuffer({ columnData: [{ name: 'string', data }], statistics: false })
-    expect(file.byteLength).toBe(178)
+    expect(file.byteLength).toBe(170)
     // round trip
     const result = await parquetReadObjects({ file })
     expect(result.length).toBe(100000)
@@ -81,8 +81,8 @@ describe('parquetWriteBuffer', () => {
   it('writes statistics when enabled', () => {
     const withStats = parquetWriteBuffer({ columnData: exampleData, statistics: true })
     const noStats = parquetWriteBuffer({ columnData: exampleData, statistics: false })
-    expect(withStats.byteLength).toBe(773)
-    expect(noStats.byteLength).toBe(663)
+    expect(withStats.byteLength).toBe(721)
+    expect(noStats.byteLength).toBe(611)
   })
 
   it('serializes list types', async () => {