Skip to content

Latest commit

 

History

History
125 lines (89 loc) · 2.8 KB

README.md

File metadata and controls

125 lines (89 loc) · 2.8 KB
Raw

PyDolphinDB

A C++ boosted DolphinDB Python3 API based on Pybind11, DolphinDB Cpp API and DolphinDB Python3 API

Core features

  • Connect to the DolphinDB server
  • Login with encrption
  • Run scripts and fetch the results
  • RPC
  • Upload supported Python objects
  • Streaming

A detailed tutorial (pydolphindb/doc/tutorial.md, still in progress) about:

  • Develop a Python C++ extension via pybind11
  • Use DolphinDB C++ API to build the C++ extension
  • Package your Python module and upload it to PyPi

Fast data conversion

pydolphindb can map the DolphinDB data structures to Python builtins

Python          DolphinDB
NoneType        VOID
bool            BOOL
int             LONG
float           DOUBLE
str             STRING
set             SET
tuple           VECTOR
list            VECTOR
dict            DICTIONARY

For efficiency and accuracy, pydolphindb support data conversions between numpy(array)/pandas(DataFrame) and DolphinDB

numpy dtype     DolphinDB
bool            BOOL
int8            CHAR
int16           SHORT
int32           INT
int64           LONG
float32         FLOAT
float64         DOUBLE
object          STRING or ANY (based on type inference)
datetime64[D]   DATE
datetime64[M]   MONTH
datetime64[ms]  TIMESTAMP
datetime64[m]   MINUTE
datetime64[s]   DATETIME
datetime64[ns]  NANOTIMESTAMP

pandas
DataFrame       TABLE

pandas does not support datetime64 otherthan datetime64[ns]

Build

1.prerequisite

  • CMake
  • Python
  • OpenSSL 1.0.2

2.start build in the pydolphindb folder

mkdir build && cd build
cmake .. -DOPENSSL_LIB_PATH=/path/to/openssl-1.0.2/lib
make && make install

Some optional macros:

-DPYTHON_EXECUTABLE=/path/to/python
-DPYTHON_MODULE_EXTENSION=extension_string
-DVERSION_INFO=version_string

3.use pydolphindb/pydolphindb as a Python module folder

installing pydolphindb via setup.py will be supported later

Tested compilers

1.GCC 4.8.5 or newer

2.Cygwin/GCC

Tested Python

Python 3.4~3.7

Python 2.7 with minor modifications

Quick start

run the following script in Python3 after installation

import pydolphindb as pydb
s = pydb.session()
isSuccess = s.connect(DOLPHINDB_IP, DOLPHINDB_PORT)
if isSuccess:
    s.run("1..100")

About

This project was created by JasonYuchen([email protected]).

License

pydolphindb is provided under MIT license that can be found in the LICENSE file.

pybind11 license

DolphinDB Cpp API license

DolphinDB Python3 API license