Engines

Iterable Data supports multiple processing engines, each optimized for different use cases.

Internal Engine (Default)

The internal engine uses pure Python implementations for all formats. It supports all file types and compression codecs.

Features

• ✅ Supports 100+ formats
• ✅ Supports all compression codecs
• ✅ Works with nested data structures
• ✅ No additional dependencies required

Usage

from iterable import open_iterable

# Internal engine is the default
# Recommended: Using context manager
with open_iterable('data.csv.gz', engine='internal') as source:
    for row in source:
        print(row)
# or simply (internal is default)
with open_iterable('data.csv.gz') as source:
    for row in source:
        print(row)

When to Use

• Working with formats not supported by DuckDB
• Need other compression codecs (BZip2, LZMA, LZ4, etc.)
• Processing nested data structures
• General-purpose data processing

DuckDB Engine

The DuckDB engine provides high-performance querying capabilities using the DuckDB database engine with advanced pushdown optimizations.

Features

• ⚡ Fast queries on large files
• ⚡ Efficient row counting
• ⚡ SQL-like operations
• ⚡ Supports compressed files
• ⚡ Column projection pushdown - Only read needed columns
• ⚡ Filter pushdown - Filter at the database level
• ⚡ Direct SQL queries - Full SQL power with iterator interface

Supported Formats

• Formats: CSV, JSONL, NDJSON, JSON, Parquet
• Compression: GZIP (.gz), ZStandard (.zst, .zstd)

Installation

DuckDB is an optional dependency:

pip install duckdb

Basic Usage

from iterable import open_iterable

# Recommended: Using context manager
# Use DuckDB engine
with open_iterable('data.csv.gz', engine='duckdb') as source:
    # Fast row counting
    total = source.totals()
    print(f"Total records: {total}")
    
    for row in source:
        print(row)
# File automatically closed

Column Projection Pushdown

Only read specific columns to reduce I/O and memory usage:

from iterable import open_iterable

# Only read 'name' and 'age' columns
with open_iterable('data.csv', engine='duckdb', 
                   iterableargs={'columns': ['name', 'age']}) as source:
    for row in source:
        print(row)  # Only contains 'name' and 'age' keys

Filter Pushdown

Filter rows at the database level for maximum performance:

from iterable import open_iterable

# SQL string filter
with open_iterable('data.csv', engine='duckdb',
                   iterableargs={'filter': "age > 18 AND status = 'active'"}) as source:
    for row in source:
        print(row)  # Only rows matching the condition

# Python callable filter (falls back to Python-side filtering if not translatable)
with open_iterable('data.csv', engine='duckdb',
                   iterableargs={'filter': lambda row: row['age'] > 18}) as source:
    for row in source:
        print(row)

Combined Pushdown

Combine column projection and filtering for maximum efficiency:

from iterable import open_iterable

# Read only 'name' and 'age' columns, filtered by age > 18
with open_iterable('data.csv', engine='duckdb',
                   iterableargs={
                       'columns': ['name', 'age'],
                       'filter': 'age > 18'
                   }) as source:
    for row in source:
        print(row)  # Only 'name' and 'age', only rows where age > 18

Direct SQL Queries

Execute full SQL queries while maintaining the iterator interface:

from iterable import open_iterable

# Custom SQL query with ORDER BY and LIMIT
with open_iterable('data.parquet', engine='duckdb',
                   iterableargs={
                       'query': 'SELECT name, age FROM read_parquet(\'data.parquet\') WHERE age > 18 ORDER BY age DESC LIMIT 100'
                   }) as source:
    for row in source:
        print(row)

# Note: When 'query' is provided, 'columns' and 'filter' parameters are ignored

Important: When using custom queries, you must reference files using DuckDB's read functions:

• CSV: read_csv_auto('file.csv')
• JSONL/JSON: read_json_auto('file.jsonl')
• Parquet: read_parquet('file.parquet')

When to Use

• ✅ Large files (100MB+)
• ✅ Need fast row counting
• ✅ Working with CSV, JSONL, JSON, or Parquet files
• ✅ Files compressed with GZIP or ZStandard
• ✅ Need SQL-like querying capabilities
• ✅ Want to reduce I/O by reading only needed columns
• ✅ Need to filter large datasets efficiently

Limitations

• ❌ Only supports CSV, JSONL, NDJSON, JSON, and Parquet formats
• ❌ Only supports GZIP and ZStandard compression
• ❌ Requires DuckDB to be installed
• ❌ Not suitable for very large streaming files
• ❌ Python callable filters fall back to Python-side filtering (may be slower)

Direct DuckDB Queries

You can also use DuckDB directly for more complex queries:

import duckdb

# Connect to DuckDB
conn = duckdb.connect()

# Query JSONL file directly
result = conn.execute("""
    SELECT * 
    FROM 'data.jsonl.zst' 
    WHERE age > 18
    LIMIT 100
""").fetchall()

for row in result:
    print(row)

Performance Comparison

Operation	Internal Engine	DuckDB Engine
Small files (< 10MB)	Fast	Fast
Large files (> 100MB)	Moderate	Very Fast
Row counting	Slow	Very Fast
Filtering	Moderate	Very Fast
Format support	All formats	CSV/JSONL only
Compression	All codecs	GZIP/ZStandard only

Choosing an Engine

Use Internal Engine when:

• Working with formats not supported by DuckDB
• Need other compression codecs
• Processing nested data structures
• General-purpose data processing

Use DuckDB Engine when:

• Working with large CSV or JSONL files
• Need fast row counting
• Need SQL-like querying
• Files are compressed with GZIP or ZStandard

Error Handling

from iterable import open_iterable

try:
    # Using DuckDB engine with error handling
    with open_iterable('data.csv.gz', engine='duckdb') as source:
        total = source.totals()
        for row in source:
            process(row)
except ImportError:
    print("DuckDB not installed. Install with: pip install duckdb")
except FileNotFoundError:
    print("File not found")
except Exception as e:
    print(f"Error with DuckDB engine: {e}")
    # Fallback to internal engine
    with open_iterable('data.csv.gz', engine='internal') as source:
        for row in source:
            process(row)

Troubleshooting

DuckDB Engine Not Available

• Install DuckDB: pip install duckdb
• Check format: DuckDB only supports CSV, JSONL, NDJSON, and JSON
• Check compression: Only GZIP and ZStandard compression supported
• Fallback: Use engine='internal' as fallback

Performance Issues

• Large files: DuckDB engine is faster for files > 100MB
• Small files: Internal engine may be faster for small files
• Format support: Use internal engine for unsupported formats
• Compression: DuckDB engine works best with GZIP and ZStandard

Database Engines

Database engines provide read-only access to SQL and NoSQL databases as iterable data sources. They enable streaming, memory-efficient access to database tables and queries.

Features

• ✅ Streaming access: Large result sets streamed in batches
• ✅ Unified interface: Works with open_iterable(), convert(), and pipeline()
• ✅ Read-only safety: All operations are read-only by default
• ✅ Memory efficient: Server-side cursors and batch processing
• ✅ Multiple databases: PostgreSQL, ClickHouse (available), MySQL, MSSQL, SQLite, MongoDB, Elasticsearch (planned)

Supported Databases

• PostgreSQL: ✅ Available (requires psycopg2-binary)
• ClickHouse: ✅ Available (requires clickhouse-connect)
• MySQL/MariaDB: 🚧 Planned
• Microsoft SQL Server: 🚧 Planned
• SQLite: 🚧 Planned
• MongoDB: 🚧 Planned
• Elasticsearch/OpenSearch: 🚧 Planned

Installation

Install database dependencies:

# All database engines
pip install iterabledata[db]

# SQL databases only
pip install iterabledata[db-sql]

# NoSQL databases only
pip install iterabledata[db-nosql]

# Specific database (PostgreSQL)
pip install psycopg2-binary

# Specific database (ClickHouse)
pip install clickhouse-connect

Basic Usage

from iterable import open_iterable

# Read from PostgreSQL database
with open_iterable(
    'postgresql://user:password@localhost:5432/mydb',
    engine='postgres',
    iterableargs={'query': 'users'}
) as source:
    for row in source:
        print(row)

# Read from ClickHouse database
with open_iterable(
    'clickhouse://user:password@localhost:9000/analytics',
    engine='clickhouse',
    iterableargs={'query': 'events'}
) as source:
    for row in source:
        print(row)

When to Use Database Engines

Use Database Engines when:

• ✅ Reading data from SQL or NoSQL databases
• ✅ Need streaming access to large database tables
• ✅ Want to use database sources with convert() and pipeline()
• ✅ Need memory-efficient processing of large result sets
• ✅ Working with ETL/analytical workloads (read-only is safe)

Use File Engines (internal/duckdb) when:

• ✅ Working with file-based data sources
• ✅ Need write operations
• ✅ Working with formats not supported by database engines
• ✅ Processing local files

Database vs File Engines

Feature	File Engines	Database Engines
Data Source	Files	Databases
Write Support	✅ Yes	❌ Read-only (write planned)
Streaming	✅ Yes	✅ Yes (server-side cursors)
Format Support	100+ formats	SQL/NoSQL databases
Memory Efficiency	Good	Excellent (batched)
Query Capabilities	Limited	Full SQL/NoSQL queries
Reset Support	✅ Yes	❌ No (queries can't be reset)

Examples

PostgreSQL

from iterable import open_iterable

# Read table
with open_iterable(
    'postgresql://localhost/mydb',
    engine='postgres',
    iterableargs={'query': 'users'}
) as source:
    for row in source:
        print(row)

# Read with filtering
with open_iterable(
    'postgresql://localhost/mydb',
    engine='postgres',
    iterableargs={
        'query': 'users',
        'columns': ['id', 'name', 'email'],
        'filter': 'active = TRUE'
    }
) as source:
    for row in source:
        print(row)

ClickHouse

from iterable import open_iterable

# Read table
with open_iterable(
    'clickhouse://localhost:9000/analytics',
    engine='clickhouse',
    iterableargs={'query': 'events'}
) as source:
    for row in source:
        print(row)

# Read with ClickHouse query settings
with open_iterable(
    'clickhouse://localhost:9000/analytics',
    engine='clickhouse',
    iterableargs={
        'query': 'large_table',
        'settings': {'max_threads': 4, 'max_memory_usage': 10000000000}
    }
) as source:
    for row in source:
        print(row)

# Read with database parameter
with open_iterable(
    'clickhouse://localhost:9000',
    engine='clickhouse',
    iterableargs={
        'query': 'events',
        'database': 'analytics',
        'columns': ['id', 'name', 'timestamp']
    }
) as source:
    for row in source:
        print(row)

Limitations

• Read-only: Write operations (INSERT, UPDATE, DELETE) are not yet supported
• No reset: Database queries cannot be reset after iteration starts
• Driver dependencies: Each database requires its own driver package

Error Handling

from iterable import open_iterable

try:
    with open_iterable(
        'postgresql://localhost/mydb',
        engine='postgres',
        iterableargs={'query': 'users'}
    ) as source:
        for row in source:
            process(row)
except ImportError:
    print("Database driver not installed. Install with: pip install psycopg2-binary")
except ConnectionError as e:
    print(f"Connection failed: {e}")

See Database Engines for comprehensive documentation.

Related Topics

• DuckDB Integration Use Case
• open_iterable() - Opening files with engines
• Database Engines - Database engine documentation
• Supported Formats