fix: stream database dumps for large databases

src/export/csv.test.ts

@@ -1,13 +1,22 @@
 import { describe, it, expect, vi, beforeEach } from 'vitest'
 import { exportTableToCsvRoute } from './csv'
-import { getTableData, createExportResponse } from './index'
+import { tableExists, getTableDataChunked } from './index'
 import { createResponse } from '../utils'
 import type { DataSource } from '../types'
 import type { StarbaseDBConfiguration } from '../handler'
 
 vi.mock('./index', () => ({
-    getTableData: vi.fn(),
-    createExportResponse: vi.fn(),
+    tableExists: vi.fn(),
+    getTableDataChunked: vi.fn(),
+    createStreamingExportResponse: vi.fn((stream, fileName, contentType) => {
+        const headers = new Headers({
+            'Content-Type': contentType,
+            'Content-Disposition': `attachment; filename="${fileName}"`,
+            'Transfer-Encoding': 'chunked',
+        })
+        return new Response(stream, { headers })
+    }),
+    CHUNK_SIZE: 5000,
 }))
 
 vi.mock('../utils', () => ({
@@ -43,46 +52,41 @@ beforeEach(() => {
 
 describe('CSV Export Module', () => {
     it('should return a CSV file when table data exists', async () => {
-        vi.mocked(getTableData).mockResolvedValue([
+        vi.mocked(tableExists).mockResolvedValue(true)
+
+        // Fewer rows than CHUNK_SIZE -> loop breaks after first call
+        vi.mocked(getTableDataChunked).mockResolvedValueOnce([
             { id: 1, name: 'Alice', age: 30 },
             { id: 2, name: 'Bob', age: 25 },
         ])
 
-        vi.mocked(createExportResponse).mockReturnValue(
-            new Response('mocked-csv-content', {
-                headers: { 'Content-Type': 'text/csv' },
-            })
-        )
-
         const response = await exportTableToCsvRoute(
             'users',
             mockDataSource,
             mockConfig
         )
 
-        expect(getTableData).toHaveBeenCalledWith(
+        expect(tableExists).toHaveBeenCalledWith(
             'users',
             mockDataSource,
             mockConfig
         )
-        expect(createExportResponse).toHaveBeenCalledWith(
-            'id,name,age\n1,Alice,30\n2,Bob,25\n',
-            'users_export.csv',
-            'text/csv'
-        )
         expect(response.headers.get('Content-Type')).toBe('text/csv')
+
+        const text = await response.text()
+        expect(text).toBe('id,name,age\n1,Alice,30\n2,Bob,25\n')
     })
 
     it('should return 404 if table does not exist', async () => {
-        vi.mocked(getTableData).mockResolvedValue(null)
+        vi.mocked(tableExists).mockResolvedValue(false)
 
         const response = await exportTableToCsvRoute(
             'non_existent_table',
             mockDataSource,
             mockConfig
         )
 
-        expect(getTableData).toHaveBeenCalledWith(
+        expect(tableExists).toHaveBeenCalledWith(
             'non_existent_table',
             mockDataSource,
             mockConfig
@@ -95,64 +99,51 @@ describe('CSV Export Module', () => {
         )
     })
 
-    it('should handle empty table (return only headers)', async () => {
-        vi.mocked(getTableData).mockResolvedValue([])
-
-        vi.mocked(createExportResponse).mockReturnValue(
-            new Response('mocked-csv-content', {
-                headers: { 'Content-Type': 'text/csv' },
-            })
-        )
+    it('should handle empty table (return empty stream)', async () => {
+        vi.mocked(tableExists).mockResolvedValue(true)
+        vi.mocked(getTableDataChunked).mockResolvedValueOnce([])
 
         const response = await exportTableToCsvRoute(
             'empty_table',
             mockDataSource,
             mockConfig
         )
 
-        expect(getTableData).toHaveBeenCalledWith(
+        expect(tableExists).toHaveBeenCalledWith(
             'empty_table',
             mockDataSource,
             mockConfig
         )
-        expect(createExportResponse).toHaveBeenCalledWith(
-            '',
-            'empty_table_export.csv',
-            'text/csv'
-        )
         expect(response.headers.get('Content-Type')).toBe('text/csv')
+
+        const text = await response.text()
+        expect(text).toBe('')
     })
 
     it('should escape commas and quotes in CSV values', async () => {
-        vi.mocked(getTableData).mockResolvedValue([
+        vi.mocked(tableExists).mockResolvedValue(true)
+
+        vi.mocked(getTableDataChunked).mockResolvedValueOnce([
             { id: 1, name: 'Sahithi, is', bio: 'my forever "penguin"' },
         ])
 
-        vi.mocked(createExportResponse).mockReturnValue(
-            new Response('mocked-csv-content', {
-                headers: { 'Content-Type': 'text/csv' },
-            })
-        )
-
         const response = await exportTableToCsvRoute(
             'special_chars',
             mockDataSource,
             mockConfig
         )
 
-        expect(createExportResponse).toHaveBeenCalledWith(
-            'id,name,bio\n1,"Sahithi, is","my forever ""penguin"""\n',
-            'special_chars_export.csv',
-            'text/csv'
+        const text = await response.text()
+        expect(text).toBe(
+            'id,name,bio\n1,"Sahithi, is","my forever ""penguin"""\n'
         )
-        expect(response.headers.get('Content-Type')).toBe('text/csv')
     })
 
     it('should return 500 on an unexpected error', async () => {
         const consoleErrorMock = vi
             .spyOn(console, 'error')
             .mockImplementation(() => {})
-        vi.mocked(getTableData).mockRejectedValue(new Error('Database Error'))
+        vi.mocked(tableExists).mockRejectedValue(new Error('Database Error'))
 
         const response = await exportTableToCsvRoute(
             'users',

src/export/csv.ts

@@ -1,51 +1,101 @@
-import { getTableData, createExportResponse } from './index'
+import {
+    tableExists,
+    getTableDataChunked,
+    createStreamingExportResponse,
+    CHUNK_SIZE,
+} from './index'
 import { createResponse } from '../utils'
 import { DataSource } from '../types'
 import { StarbaseDBConfiguration } from '../handler'
 
+/**
+ * Format a single value for CSV output.
+ * Wraps in double quotes and escapes inner quotes when the value contains
+ * commas, double quotes, or newlines.
+ */
+function formatCsvValue(value: unknown): string {
+    if (value === null || value === undefined) {
+        return ''
+    }
+
+    const str = String(value)
+    if (str.includes(',') || str.includes('"') || str.includes('\n')) {
+        return `"${str.replace(/"/g, '""')}"`
+    }
+    return str
+}
+
 export async function exportTableToCsvRoute(
     tableName: string,
     dataSource: DataSource,
     config: StarbaseDBConfiguration
 ): Promise<Response> {
     try {
-        const data = await getTableData(tableName, dataSource, config)
+        const exists = await tableExists(tableName, dataSource, config)
 
-        if (data === null) {
+        if (!exists) {
             return createResponse(
                 undefined,
                 `Table '${tableName}' does not exist.`,
                 404
             )
         }
 
-        // Convert the result to CSV
-        let csvContent = ''
-        if (data.length > 0) {
-            // Add headers
-            csvContent += Object.keys(data[0]).join(',') + '\n'
-
-            // Add data rows
-            data.forEach((row: any) => {
-                csvContent +=
-                    Object.values(row)
-                        .map((value) => {
-                            if (
-                                typeof value === 'string' &&
-                                (value.includes(',') ||
-                                    value.includes('"') ||
-                                    value.includes('\n'))
-                            ) {
-                                return `"${value.replace(/"/g, '""')}"`
-                            }
-                            return value
-                        })
-                        .join(',') + '\n'
-            })
-        }
+        const encoder = new TextEncoder()
+        let headerWritten = false
+
+        const stream = new ReadableStream({
+            async start(controller) {
+                try {
+                    let offset = 0
+
+                    while (true) {
+                        const rows = await getTableDataChunked(
+                            tableName,
+                            offset,
+                            CHUNK_SIZE,
+                            dataSource,
+                            config
+                        )
+
+                        if (!rows || rows.length === 0) {
+                            break
+                        }
+
+                        let chunk = ''
+
+                        // Write header row from first batch of results
+                        if (!headerWritten) {
+                            chunk += Object.keys(rows[0]).join(',') + '\n'
+                            headerWritten = true
+                        }
+
+                        // Write data rows
+                        for (const row of rows) {
+                            chunk +=
+                                Object.values(row)
+                                    .map(formatCsvValue)
+                                    .join(',') + '\n'
+                        }
+
+                        controller.enqueue(encoder.encode(chunk))
+
+                        if (rows.length < CHUNK_SIZE) {
+                            break
+                        }
+
+                        offset += CHUNK_SIZE
+                    }
+
+                    controller.close()
+                } catch (error) {
+                    controller.error(error)
+                }
+            },
+        })
 
-        return createExportResponse(
-            csvContent,
+        return createStreamingExportResponse(
+            stream,
             `${tableName}_export.csv`,
             'text/csv'
         )