chdb 3.4.0

chDB is an in-process SQL OLAP Engine powered by ClickHouse

📢 chDB joins the ClickHouse family 🐍+🚀

chDB

chDB is an in-process SQL OLAP Engine powered by ClickHouse [^1] For more details: The birth of chDB

Features

• In-process SQL OLAP Engine, powered by ClickHouse
• No need to install ClickHouse
• Minimized data copy from C++ to Python with python memoryview
• Input&Output support Parquet, CSV, JSON, Arrow, ORC and 60+more formats, samples
• Support Python DB API 2.0, example

Arch

Get Started

Get started with chdb using our Installation and Usage Examples

Installation

Currently, chDB supports Python 3.8+ on macOS and Linux (x86_64 and ARM64).

pip install chdb

Usage

Run in command line

python3 -m chdb SQL [OutputFormat]

python3 -m chdb "SELECT 1,'abc'" Pretty

Data Input

The following methods are available to access on-disk and in-memory data formats:

🗂️ Connection based API (recommended)

import chdb

# Create a connection (in-memory by default)
conn = chdb.connect(":memory:")
# Or use file-based: conn = chdb.connect("test.db")

# Create a cursor
cur = conn.cursor()

# Execute queries
cur.execute("SELECT number, toString(number) as str FROM system.numbers LIMIT 3")

# Fetch data in different ways
print(cur.fetchone())    # Single row: (0, '0')
print(cur.fetchmany(2))  # Multiple rows: ((1, '1'), (2, '2'))

# Get column information
print(cur.column_names())  # ['number', 'str']
print(cur.column_types())  # ['UInt64', 'String']

# Use the cursor as an iterator
cur.execute("SELECT number FROM system.numbers LIMIT 3")
for row in cur:
    print(row)

# Always close resources when done
cur.close()
conn.close()

For more details, see examples/connect.py.

🗂️ Query On File

(Parquet, CSV, JSON, Arrow, ORC and 60+)

You can execute SQL and return desired format data.

import chdb
res = chdb.query('select version()', 'Pretty'); print(res)

Work with Parquet or CSV

# See more data type format in tests/format_output.py
res = chdb.query('select * from file("data.parquet", Parquet)', 'JSON'); print(res)
res = chdb.query('select * from file("data.csv", CSV)', 'CSV');  print(res)
print(f"SQL read {res.rows_read()} rows, {res.bytes_read()} bytes, storage read {res.storage_rows_read()} rows, {res.storage_bytes_read()} bytes, elapsed {res.elapsed()} seconds")

Pandas dataframe output

# See more in https://clickhouse.com/docs/en/interfaces/formats
chdb.query('select * from file("data.parquet", Parquet)', 'Dataframe')

🗂️ Query On Table

(Pandas DataFrame, Parquet file/bytes, Arrow bytes)

Query On Pandas DataFrame

import chdb.dataframe as cdf
import pandas as pd
# Join 2 DataFrames
df1 = pd.DataFrame({'a': [1, 2, 3], 'b': ["one", "two", "three"]})
df2 = pd.DataFrame({'c': [1, 2, 3], 'd': ["①", "②", "③"]})
ret_tbl = cdf.query(sql="select * from __tbl1__ t1 join __tbl2__ t2 on t1.a = t2.c",
                  tbl1=df1, tbl2=df2)
print(ret_tbl)
# Query on the DataFrame Table
print(ret_tbl.query('select b, sum(a) from __table__ group by b'))
# Pandas DataFrames are automatically registered as temporary tables in ClickHouse
chdb.query("SELECT * FROM Python(df1) t1 JOIN Python(df2) t2 ON t1.a = t2.c").show()

🗂️ Query with Stateful Session

from chdb import session as chs

## Create DB, Table, View in temp session, auto cleanup when session is deleted.
sess = chs.Session()
sess.query("CREATE DATABASE IF NOT EXISTS db_xxx ENGINE = Atomic")
sess.query("CREATE TABLE IF NOT EXISTS db_xxx.log_table_xxx (x String, y Int) ENGINE = Log;")
sess.query("INSERT INTO db_xxx.log_table_xxx VALUES ('a', 1), ('b', 3), ('c', 2), ('d', 5);")
sess.query(
    "CREATE VIEW db_xxx.view_xxx AS SELECT * FROM db_xxx.log_table_xxx LIMIT 4;"
)
print("Select from view:\n")
print(sess.query("SELECT * FROM db_xxx.view_xxx", "Pretty"))

see also: test_stateful.py.

🗂️ Query with Python DB-API 2.0

import chdb.dbapi as dbapi
print("chdb driver version: {0}".format(dbapi.get_client_info()))

conn1 = dbapi.connect()
cur1 = conn1.cursor()
cur1.execute('select version()')
print("description: ", cur1.description)
print("data: ", cur1.fetchone())
cur1.close()
conn1.close()

🗂️ Query with UDF (User Defined Functions)

from chdb.udf import chdb_udf
from chdb import query

@chdb_udf()
def sum_udf(lhs, rhs):
    return int(lhs) + int(rhs)

print(query("select sum_udf(12,22)"))

Some notes on chDB Python UDF(User Defined Function) decorator.

1. The function should be stateless. So, only UDFs are supported, not UDAFs(User Defined Aggregation Function).
2. Default return type is String. If you want to change the return type, you can pass in the return type as an argument. The return type should be one of the following: https://clickhouse.com/docs/en/sql-reference/data-types
3. The function should take in arguments of type String. As the input is TabSeparated, all arguments are strings.
4. The function will be called for each line of input. Something like this:

   def sum_udf(lhs, rhs):
       return int(lhs) + int(rhs)
   
   for line in sys.stdin:
       args = line.strip().split('\t')
       lhs = args[0]
       rhs = args[1]
       print(sum_udf(lhs, rhs))
       sys.stdout.flush()
   

5. The function should be pure python function. You SHOULD import all python modules used IN THE FUNCTION.

   def func_use_json(arg):
       import json
       ...
   

6. Python interpertor used is the same as the one used to run the script. Get from sys.executable

see also: test_udf.py.

🗂️ Streaming Query

Process large datasets with constant memory usage through chunked streaming.

from chdb import session as chs

sess = chs.Session()

# Example 1: Basic example of using streaming query
rows_cnt = 0
with sess.send_query("SELECT * FROM numbers(200000)", "CSV") as stream_result:
    for chunk in stream_result:
        rows_cnt += chunk.rows_read()

print(rows_cnt) # 200000

# Example 2: Manual iteration with fetch()
rows_cnt = 0
stream_result = sess.send_query("SELECT * FROM numbers(200000)", "CSV")
while True:
    chunk = stream_result.fetch()
    if chunk is None:
        break
    rows_cnt += chunk.rows_read()

print(rows_cnt) # 200000

# Example 3: Early cancellation demo
rows_cnt = 0
stream_result = sess.send_query("SELECT * FROM numbers(200000)", "CSV")
while True:
    chunk = stream_result.fetch()
    if chunk is None:
        break
    if rows_cnt > 0:
        stream_result.cancel()
        break
    rows_cnt += chunk.rows_read()

print(rows_cnt) # 65409

sess.close()

For more details, see test_streaming_query.py.

🗂️ Python Table Engine

Query on Pandas DataFrame

import chdb
import pandas as pd
df = pd.DataFrame(
    {
        "a": [1, 2, 3, 4, 5, 6],
        "b": ["tom", "jerry", "auxten", "tom", "jerry", "auxten"],
        "dict_col": [
            {'id': 1, 'tags': ['urgent', 'important'], 'metadata': {'created': '2024-01-01'}},
            {'id': 2, 'tags': ['normal'], 'metadata': {'created': '2024-02-01'}},
            {'id': 3, 'name': 'tom'},
            {'id': 4, 'value': '100'},
            {'id': 5, 'value': 101},
            {'id': 6, 'value': 102},
        ],
    }
)

chdb.query("SELECT b, sum(a) FROM Python(df) GROUP BY b ORDER BY b").show()
chdb.query("SELECT dict_col.id FROM Python(df) WHERE dict_col.value='100'").show()

Query on Arrow Table

import chdb
import pyarrow as pa
arrow_table = pa.table(
    {
        "a": [1, 2, 3, 4, 5, 6],
        "b": ["tom", "jerry", "auxten", "tom", "jerry", "auxten"],
        "dict_col": [
            {'id': 1, 'value': 'tom'},
            {'id': 2, 'value': 'jerry'},
            {'id': 3, 'value': 'auxten'},
            {'id': 4, 'value': 'tom'},
            {'id': 5, 'value': 'jerry'},
            {'id': 6, 'value': 'auxten'},
        ],
    }
)

chdb.query("SELECT b, sum(a) FROM Python(arrow_table) GROUP BY b ORDER BY b").show()
chdb.query("SELECT dict_col.id FROM Python(arrow_table) WHERE dict_col.value='tom'").show()

Query on chdb.PyReader class instance

1. You must inherit from chdb.PyReader class and implement the read method.
2. The read method should:
   1. return a list of lists, the first demension is the column, the second dimension is the row, the columns order should be the same as the first arg col_names of read.
   2. return an empty list when there is no more data to read.
   3. be stateful, the cursor should be updated in the read method.
3. An optional get_schema method can be implemented to return the schema of the table. The prototype is def get_schema(self) -> List[Tuple[str, str]]:, the return value is a list of tuples, each tuple contains the column name and the column type. The column type should be one of the following: https://clickhouse.com/docs/en/sql-reference/data-types

import chdb

class myReader(chdb.PyReader):
    def __init__(self, data):
        self.data = data
        self.cursor = 0
        super().__init__(data)

    def read(self, col_names, count):
        print("Python func read", col_names, count, self.cursor)
        if self.cursor >= len(self.data["a"]):
            self.cursor = 0
            return []
        block = [self.data[col] for col in col_names]
        self.cursor += len(block[0])
        return block

    def get_schema(self):
        return [
            ("a", "int"),
            ("b", "str"),
            ("dict_col", "json")
        ]

reader = myReader(
    {
        "a": [1, 2, 3, 4, 5, 6],
        "b": ["tom", "jerry", "auxten", "tom", "jerry", "auxten"],
        "dict_col": [
            {'id': 1, 'tags': ['urgent', 'important'], 'metadata': {'created': '2024-01-01'}},
            {'id': 2, 'tags': ['normal'], 'metadata': {'created': '2024-02-01'}},
            {'id': 3, 'name': 'tom'},
            {'id': 4, 'value': '100'},
            {'id': 5, 'value': 101},
            {'id': 6, 'value': 102}
        ],
    }
)

chdb.query("SELECT b, sum(a) FROM Python(reader) GROUP BY b ORDER BY b").show()
chdb.query("SELECT dict_col.id FROM Python(reader) WHERE dict_col.value='100'").show()

see also: test_query_py.py and test_query_json.py.

JSON Type Inference

chDB automatically converts Python dictionary objects to ClickHouse JSON types from these sources:

1. Pandas DataFrame

   • Columns with object dtype are sampled (default 10,000 rows) to detect JSON structures.
   • Control sampling via SQL settings:

     SET pandas_analyze_sample = 10000  -- Default sampling
     SET pandas_analyze_sample = 0      -- Force String type
     SET pandas_analyze_sample = -1     -- Force JSON type
     

   • Columns are converted to String if sampling finds non-dictionary values.

2. Arrow Table

   • struct type columns are automatically mapped to JSON columns.
   • Nested structures preserve type information.

3. chdb.PyReader

   • Implement custom schema mapping in get_schema():

     def get_schema(self):
         return [
             ("c1", "JSON"),  # Explicit JSON mapping
             ("c2", "String")
         ]
     

   • Column types declared as "JSON" will bypass auto-detection.

When converting Python dictionary objects to JSON columns:

1. Nested Structures

   • Recursively process nested dictionaries, lists, tuples and NumPy arrays.

2. Primitive Types

   • Automatic type recognition for basic types such as integers, floats, strings, and booleans, and more.

3. Complex Objects

   • Non-primitive types will be converted to strings.

Limitations

1. Column types supported: pandas.Series, pyarrow.array, chdb.PyReader
2. Data types supported: Int, UInt, Float, String, Date, DateTime, Decimal
3. Python Object type will be converted to String
4. Pandas DataFrame performance is all of the best, Arrow Table is better than PyReader

For more examples, see examples and tests.

Demos and Examples

• Project Documentation and Usage Examples
• Colab Notebooks and other Script Examples

Benchmark

• ClickBench of embedded engines

• chDB vs Pandas

• Benchmark on DataFrame: chDB Pandas DuckDB Polars

Documentation

• For chdb specific examples and documentation refer to chDB docs
• For SQL syntax, please refer to ClickHouse SQL Reference

Events

• Demo chDB at ClickHouse v23.7 livehouse! and Slides

Contributing

Contributions are what make the open source community such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated. There are something you can help:

• Help test and report bugs
• Help improve documentation
• Help improve code quality and performance

Bindings

We welcome bindings for other languages, please refer to bindings for more details.

Paper

• ClickHouse - Lightning Fast Analytics for Everyone

License

Apache 2.0, see LICENSE for more information.

Acknowledgments

chDB is mainly based on ClickHouse [^1] for trade mark and other reasons, I named it chDB.

Contact

• Discord: https://discord.gg/D2Daa2fM5K
• Email: [e-mail]
• Twitter: @chdb

[^1]: ClickHouse® is a trademark of ClickHouse Inc. All trademarks, service marks, and logos mentioned or depicted are the property of their respective owners. The use of any third-party trademarks, brand names, product names, and company names does not imply endorsement, affiliation, or association with the respective owners.