DataStore: Pandas-compatible API with SQL optimization

DataStore is chDB’s pandas-compatible API that combines the familiar pandas DataFrame interface with the power of SQL query optimization and allows you to write pandas-style code while getting ClickHouse performance.

Key features

• Pandas Compatibility: 209 pandas DataFrame methods, 56 .str methods, 42+ .dt methods
• SQL Optimization: Operations automatically compile to optimized SQL queries
• Lazy Evaluation: Operations are deferred until results are needed
• 630+ API Methods: Comprehensive API surface for data manipulation
• ClickHouse Extensions: Additional accessors (.arr, .json, .url, .ip, .geo) not available in pandas

Architecture

DataStore uses lazy evaluation with dual-engine execution:

1. Lazy Operation Chain: Operations are recorded, not executed immediately
2. Smart Engine Selection: QueryPlanner routes each segment to optimal engine (chDB for SQL, Pandas for complex ops)
3. Intermediate Caching: Results cached at each step for fast iterative exploration

See Execution Model for details.

One-Line migration from Pandas

# Before (pandas)
import pandas as pd

df = pd.read_csv("data.csv")
result = df[df['age'] > 25].groupby('city')['salary'].mean()

# After (DataStore) - just change the import!
from chdb import datastore as pd
df = pd.read_csv("data.csv")
result = df[df['age'] > 25].groupby('city')['salary'].mean()

Your existing pandas code works unchanged, but now runs on the ClickHouse engine.

Performance comparison

DataStore delivers significant performance improvements over pandas, especially for aggregation and complex pipelines:

Benchmark on 10M rows. See benchmark script and Performance Guide for details.

When to use DataStore

Use DataStore when:

• Working with large datasets (millions of rows)
• Performing aggregations and groupby operations
• Querying data from files, databases, or cloud storage
• Building complex data pipelines
• You want pandas API with better performance

Use raw SQL API when:

• You prefer writing SQL directly
• You need fine-grained control over query execution
• Working with ClickHouse-specific features not exposed in pandas API

Feature comparison

API statistics

Documentation navigation

Getting Started

• Quickstart - Installation and basic usage
• Migration from Pandas - Step-by-step migration guide

API reference

• Factory Methods - Creating DataStore from various sources
• Query Building - SQL-style query operations
• Pandas Compatibility - All 209 pandas-compatible methods
• Accessors - String, DateTime, Array, JSON, URL, IP, Geo accessors
• Aggregation - Aggregate and window functions
• I/O Operations - Reading and writing data

Advanced topics

• Execution Model - Lazy evaluation and caching
• Class Reference - Complete API reference

Configuration & debugging

• Configuration - All configuration options
• Performance Mode - SQL-first mode for maximum throughput
• Debugging - Explain, profiling, and logging

Pandas user guides

• Pandas Cookbook - Common patterns
• Key Differences - Important differences from pandas
• Performance Guide - Optimization tips
• SQL for Pandas Users - Understanding the SQL behind pandas operations

Quick example

from chdb import datastore as pd

# Read data from various sources
ds = pd.read_csv("sales.csv")
# or: ds = pd.DataStore.uri("s3://bucket/sales.parquet")
# or: ds = pd.DataStore.from_mysql("mysql://user:pass@host/db/table")

# Familiar pandas operations - automatically optimized to SQL
result = (ds
    .filter(ds['amount'] > 1000)           # WHERE amount > 1000
    .groupby('region')                      # GROUP BY region
    .agg({'amount': ['sum', 'mean']})       # SUM(amount), AVG(amount)
    .sort_values('sum', ascending=False)    # ORDER BY sum DESC
    .head(10)                               # LIMIT 10
)

# View the generated SQL
print(result.to_sql())

# Execute and get results
df = result.to_df()  # Returns pandas DataFrame

Next steps

• New to DataStore? Start with the Quickstart Guide
• Coming from pandas? Read the Migration Guide
• Want to learn more? Explore the API Reference