Accessing data is the first step in using the tools described in this book . I will highlight pandas Data input and output , Although there are many tools for this purpose in other libraries .

Input and output can be divided into several categories ： Read text files and other more efficient disk storage formats , Load the data in the database , utilize Web API Operate network resources .

6.1 Read and write data in text format

pandas Provides some methods for reading tabular data as DataFrame Object function . surface 6-1 They are summarized , among read_csv and read_table It may be the one you will use most in the future .

I'll give you an overview of how these functions convert text data to DataFrame Some of the techniques used in . The options of these functions can be divided into the following categories ：

• Indexes ： Treat one or more columns as returned DataFrame Handle , And whether from the file 、 User get column name .
• Type inference and data conversion ： Including conversion of user-defined values 、 And custom missing value tag list .
• Date resolution ： Including combined functions , For example, combine the date and time information scattered in multiple columns into a single column in the result .
• iteration ： Supports block by block iteration of large files .
• Irregular data problem ： Skip some lines 、 The footer 、 Notes or other unimportant things （ For example, numerical data separated by thousands of commas ）.

Because the actual data encountered in the work may be very confusing , Some data loading functions （ In especial read_csv） The options for are becoming increasingly complex . Facing different parameters , It's normal to feel headache （read_csv There are more than 50 Parameters ）.pandas The documentation has examples of these parameters , If you find it difficult to read a document , You can use enough similar examples