python pandas 對各類文件的讀寫 IO tools

The pandas I/O API is a set of top level reader functions accessed like pd.read_csv() that generally return a pandasobject.

• read_csv
• read_excel
• read_hdf
• read_sql
• read_json
• read_msgpack (experimental)
• read_html
• read_gbq (experimental)
• read_stata
• read_sas
• read_clipboard
• read_pickle

The corresponding writer functions are object methods that are accessed like df.to_csv()

• to_csv
• to_excel
• to_hdf
• to_sql
• to_json
• to_msgpack (experimental)
• to_html
• to_gbq (experimental)
• to_stata
• to_clipboard
• to_pickle

Here is an informal performance comparison for some of these IO methods.

 

http://pandas.pydata.org/pandas-docs/stable/io.html