A brief introduction to Python's Pandas library

Pandas is the data analysis processing library for Python
Import Pandas as PD

1. read CSV, TXT file

Foodinfo = Pd.read_csv ("pandas_study.csv""utf-8")

2, view the first n, after n information

Foodinfo.head (n) foodinfo.tail (n)

3, check the format of the data frame, is dataframe or Ndarray

Print (Type (foodinfo)) # results: <class ' pandas.core.frame.DataFrame ' >

4. See what columns are available

Foodinfo.columns

5, see a few rows of several columns

Foodinfo.shape

6. Print a line, a few rows of data

foodinfo.loc[0]foodinfo.loc[0:2]foodinfo.loc[[2, 5, ten]]    # Note that the inside is an array

7. Print a column, a few columns of data

foodinfo["dti"]foodinfo[["int_rate " " DTI "]]    # Note that the inside is an array # or:columns = ["int_rate""dti"] Foodinfo[columns]

8. Print data types for all columns

Foodinfo.dtypes

9, some related operations on the column

Col_columns == [] for in col_columns:if c.endswith (" s " ): New_columus.append (c)print(c) Foodinfo[new_columus]

10, Subtraction: Multiply each line by 100 (subtraction same)

foodinfo[["int_rate""dti"] * 100

11. Add a column

New_col = foodinfo["int_rate"] *foodinfo["new_col "] = New_col

12. Operations between columns

foodinfo["dti"] * foodinfo["int_rate"]

13. View the maximum, minimum, and average values for a column

foodinfo["int_rate"].max () foodinfo["int_rate"  ].min () foodinfo["int_rate"].mean ()

14. Sort by a field-ascending

# InPlace whether to create a new dataframe,true does not require foodinfo.sort_values ("int_rate_one", inplace = True) # Sort by a field-descending foodinfo.sort_values ("int_rate_one", inplace = True, ascending = False)

15. View some properties of the data frame: maximum, minimum, mean, four-digit, etc.

Foodinfo.describe ()

16. Null Value related Operations

Pin = foodinfo["pin"#  See all empty values #  Find all empty lines len (pin_isnull_list)        # number of NULL values

17, missing value related operations

# The simple approach is to filter out null values books = foodinfo["life_cycle_books"== foodinfo["life_cycle_books"][book_isnull= = SUM (book_list_isnull)/Len ( Book_list_isnull)    #  calculate Average

18, according to the criteria to print a column of data

foodinfo[foodinfo["life_cycle_books"] = = 1]

19. Pivot Table

Import NumPy as NP # Index: The column to pivot # values: The relationship columns to compare # Aggfunc: Specific relationship, default value: Np.meandata_foodinfo = foodinfo.pivot_table (index = ["life_cycle_books  "potential_value_books" "risk_level ", Aggfunc = np.mean)print(data_foodinfo)

20. Delete missing values

# All lines na_foodinfo = Foodinfo.dropna (axis = 1)#  You can specify the column na_foodinfo = Foodinfo.dropna ( Axis = 0, subset = ["life_cycle_books" "potential_value_books" ])

21, free to take data such as: Take 80 rows life_cycle_books column

" Life_cycle_books "]

22. Re-rank Index

Foodinfo.reset_index (drop = True)

23. Custom Function: Returns the number of empty values

def count_null_columns (column):     = pd.isnull (column)    = Column[column_null]    = len (list_null)    return count_nullfoodinfo.apply (count_null_columns)

24. Series

# pandas three types of data structures # Series # DataFrame # Panel  from Import Series

25. Series shows a column of data

Series_name = taitan["name"]series_name.values

26. Positioning a row of a column

Series_name = taitan["name"= taitan["Age"  = Series (series_age.values, index = series_name) series_custom[["Ahlin, Mrs Johan (Johanna Persdotter Larsson)""asplund, Mrs Carl Oscar (Selma Augusta Emilia Johansson) c14> "]#  Description: series_custom[" "] by Column series_custom[[" "]] by row

27, take 5-10 rows of data, and the same as above:

SERIES_CUSTOM[5:10]

28. Index Transformation

Old_index == = Series_custom.reindex (sort_index)print(new_index)

29. A series of functions sorted by index and value

SC1 = series_custom.sort_index ()print= series_custom.sort_values ()Print (SC2)

30. Series Filter

Series_custom > 0.5> 0.5> 0.5) & (Series_custom < 0.9)]#  Note: &, | They're all single symbols .

31, DataFrame

# Series is a row of data, Dataframe is multiple rows of data # DataFrame can be seen as df = pd.read_csv ("titanic_train.csv") consisting of multiple Series

32. Index transformation of Dataframe

# whether the drop creates a new Df,true false Yes (indicates that a column of name is also retained, otherwise it cannot be evaluated)Df_name = Df.set_index ("name" , drop = False)

33. Dataframe view data of a certain type

types ="float64"].indexdf_name[float_columns]

34, Dataframe to find the variance

FLOAT_DF = df_name[float_columns]float_df.apply (Lambda x:np.std (x))

A brief introduction to Python's Pandas library