[Python] Summary of table creation method using DataFrame (pandas)

There are multiple ways to create a single table using Pandas DataFrame.

Summary of basic method of making.

The fastest way to read the base table is to create it as a csv file.

**table of contents**

[Table to be created](# 1 Table to be created)
[Create from 2D list](# 22 Create from 2D list)
[Specify matrix name later](# 1 Matrix name specified later)
[Specify as an option when creating a table](# 2 Specify as an option when creating a table)
[Specify the column that will be the row name from the table](# 3 Specify the column that will be the row name from the table)
[Create from 1D list](Create from # 31D list)
[Combine each list together](# 1 Combine each list together)
[Combine each list one by one](# 2 Join each list one by one)
[Merge with formula](# 3 Combine with formula)
[Create and read with csv file](# 4 Create and read with csv file)
[List of methods used](# 5 List of methods used)

## 1. 1. Table to create Create the table below.

--Column name: row0 ~ row5 --Line name: col0 ~ col4

col0：1~6
col1：col0*100
col2：col0/3
col3：AAA～FFF
col4：col3 & col1

## 2. Created from a 2D list ** ■ What is a two-dimensional array? ** `[['A','B',,],['C','D',,],,,]`

A one-dimensional list ['A','B' ,,] in the same [] in parallel.

Check with the following 3 patterns. (1) Specify the matrix name later (2) Optional when creating a table (3) Specify the column that will be the row name from the table

### (1) Specify the matrix name later

** ▼ Procedure ** ① Make each line one list ② Convert to table ③ Change the line name ④ Change column name

** ① Arrange each line in parallel as one list **

`Two-dimensional array`


list = [
[1, 100, 0.33, 'AAA', 'AAA100'], 
[2, 200, 0.67, 'BBB', 'BBB200'], 
[3, 300, 1, 'CCC', 'CCC300'], 
[4, 400, 1.33, 'DDD', 'DDD400'], 
[5, 500, 1.67, 'EEE', 'EEE500'], 
[6, 600, 2, 'FFF', 'FFF600']]

** ② Convert to table ** `pd.DataFrame(list)` └ Convert to a table with DataFrame of pandas.

`Convert to table`


df = pd.DataFrame(list)
df

Row and column names will be automatically assigned numbers.

** ③ Change line name ** `df.index = ['A','B',,,]`

-Change the line name (index) to the specified name. -Match the number of ** elements in [] with the number of rows **. Otherwise an error.

`Rename line`


df.index = ['row0','row1','row2','row3','row4','row5']
df

** ④ Change column name ** `df.columns = ['a','b',,,]`

-Change the column name to the specified name. -Match the number of ** elements in [] with the number of rows **. Otherwise an error.

`Rename column`


df.columns = ['col0','col1','col2','col3','col4']
df

This completes.

### (2) Optional when creating a table

** ▼ Procedure ** ① Make each line one list (2) Convert to table (specify matrix name as an option) 2-1 columns option 2-2 index option

** ① Arrange each line in parallel as one list **

`Two-dimensional array`


list = [
[1, 100, 0.33, 'AAA', 'AAA100'], 
[2, 200, 0.67, 'BBB', 'BBB200'], 
[3, 300, 1, 'CCC', 'CCC300'], 
[4, 400, 1.33, 'DDD', 'DDD400'], 
[5, 500, 1.67, 'EEE', 'EEE500'], 
[6, 600, 2, 'FFF', 'FFF600']]

** ② Convert to table (specify matrix name as an option) ** `pd.DataFrame(list, index=['A','B',,,], columns=['a','b',,,])` └ "list": Array to be tabulated └ index = ['A','B' ,,,]: index option └ columns = ['a','b' ,,,]: columns option └ Variables can be specified in the option [].

`Optional matrix name`


ind = ['row0','row1','row2','row3','row4','row5']
col = ['col0','col1','col2','col3','col4']

df = pd.DataFrame(list, index=ind, columns=col)
df

This completes.

### (3) Specify the column that will be the row name from the table

** ▼ Procedure ** ① Make each line one list ② Convert to table ③ Change the column name ④ Specify the column that will be the row name

** ① Make each line into one list ** Include row names in a two-dimensional array.

`Two-dimensional array`


list = [
['row0', 1, 100, 0.33, 'AAA', 'AAA100'], 
['row1', 2, 200, 0.67, 'BBB', 'BBB200'], 
['row2', 3, 300, 1, 'CCC', 'CCC300'], 
['row3', 4, 400, 1.33, 'DDD', 'DDD400'], 
['row4', 5, 500, 1.67, 'EEE', 'EEE500'], 
['row5', 6, 600, 2, 'FFF', 'FFF600'], ]

** ② Convert to table ** `pd.DataFrame(list)` └ "list": Variable containing array data

`Convert to table`


df = pd.DataFrame(list)
df

Row and column names will be automatically assigned numbers.

** ③ Change column name ** `df.columns = ['a','b',,,]`

-Change the column name to the specified name. -Match the number of ** elements in [] with the number of rows **. Otherwise an error.

In order to prevent multi-index by specifying the row that will be the next column name (set_index), leave the column name that will be the row name blank "''".

`Rename column`


df.columns =  ['','col0','col1','col2','col3','col4']
df

** ④ Specify the column that will be the row name ** `set_index ('column name')`

`Specify the column that will be the row name`


df = df.set_index('')
df

The column name "''" was set in the heading.

## 3.1 Created from a one-dimensional list

Check with the following 3 patterns. (1) Combine each list together (2) Combine each list one by one (3) Combine with mathematical formula

### (1) Combine each list together

** ▼ Procedure ** ① Create a list for each line ② Convert to a table at once ③ Change the line name ④ Change column name

** ① Create a list for each line **

`create list`


listA = [1, 100, 0.33, 'AAA', 'AAA100']
listB = [2, 200, 0.67, 'BBB', 'BBB200']
listC = [3, 300, 1, 'CCC', 'CCC300']
listD = [4, 400, 1.33, 'DDD', 'DDD400']
listE = [5, 500, 1.67, 'EEE', 'EEE500']
listF = [6, 600, 2, 'FFF', 'FFF600']

** ② Convert all together to a table ** `pd.DataFrame([A,B,,,])` └ "A" "B" one-dimensional list └ The number of elements is equal

`Convert to a table at once`


pd.DataFrame([listA,listB,listC,listD,listE,listF])

** ③ Change line name ** `df.index = ['A','B',,,]`

-Change the line name (index) to the specified name. -Match the number of ** elements in [] with the number of rows **. Otherwise an error.

`Rename line`


df.index = ['row0','row1','row2','row3','row4','row5']
df

** ④ Change column name ** `df.columns = ['a','b',,,]`

-Change the column names to the specified name. -Match the number of ** elements in [] with the number of rows **. Otherwise an error.

`Rename column`


df.columns = ['col0','col1','col2','col3','col4']
df

This completes.

### (2) Combine each list one by one

** ▼ Procedure ** ① Create a list for each line ② Create a table ③ Add a column to the table ④ Change the line name ⑤ Transpose

** ① Create a list for each line **

`create list`


listA = [1, 100, 0.33, 'AAA', 'AAA100']
listB = [2, 200, 0.67, 'BBB', 'BBB200']
listC = [3, 300, 1, 'CCC', 'CCC300']
listD = [4, 400, 1.33, 'DDD', 'DDD400']
listE = [5, 500, 1.67, 'EEE', 'EEE500']
listF = [6, 600, 2, 'FFF', 'FFF600']

** ② Creating a table ** Generate a table from one list. `pd.DataFrame(list, columns=['A'])` └ "listA": Array data └ "columns = ['A']": Set the column name to A

`Creating a table`


df= pd.DataFrame(listA, columns=['row0'])
df

When reading one column, the matrix is transposed.
** ▼ Supplement: How to create a table with column names ** Another way to specify column names when creating a table. pd.DataFrame({'A':[1,2,,],'B':[3,4,,],})

└ Describe 'column name': [array] in ({})

`Create with column name ①`


pd.DataFrame({'row0':[1, 100, 0.33, 'AAA', 'AAA100']})

`Create with column name (multiple columns)`


pd.DataFrame({'row0':[1, 100, 0.33, 'AAA', 'AAA100'], 'row1':[2, 200, 0.67, 'BBB', 'BBB200']})

** ③ Add column to table ** `df['B'] = pd.DataFrame(list)` └ "B": Column name to add └ "list": Table data to be added

`Add column to table`


df['row1'] = pd.DataFrame(listB)
df['row2'] = pd.DataFrame(listC)
df['row3'] = pd.DataFrame(listD)
df['row4'] = pd.DataFrame(listE)
df['row5'] = pd.DataFrame(listF)
df

** ④ Change line name ** `df.index = ['A','B',,,]`

`Rename line`


df.index = ['col0','col1','col2','col3','col4']
df

** ⑤ Transpose ** `df.T` └ Swap the matrix.

`Transpose`


df = df.T
df

This completes.

### (3) Combine with mathematical formula If each column has regularity, you can use mathematical formulas to join the columns.

** ▼ Regularity **

col1：col0*100
col2：col0/3
col4：col3 & col1

** ▼ Procedure ** ① Create list ② Create a table ③ Add columns to the table using formulas ④ Change the line name

** ① Create list ** Create a base column.

Two for col0 and col3.

`create list`


listA = [1,2,3,4,5,6]
listD = ['AAA', 'BBB', 'CCC', 'DDD', 'EEE', 'FFF']

** ② Creating a table ** `pd.DataFrame(list, columns=['A'])` └ "listA": Array data └ "columns = ['A']": Set the column name to A

`Creating a table`


df = pd.DataFrame(listA, columns=['col0'])
df

** ③ Add columns to the table using formulas ** ・ `Df ['B'] = df ['A'] / 2` └ "B": Column name to add └ "df ['A']" / 2: Substitute the value of the existing column name A divided by 2.

・ Df ['C'] = df ['A'] + df ['B'] └ Combine the values in columns "A" and "B" └ * str + int is an error (match the type)

・ Df ['A']. Astype (str) └ Change column "A" to "str" type

`Add columns to the table using formulas`


df['col1'] = df['col0'] * 100
df['col2'] = df['col0'] / 3
df['col3'] = pd.DataFrame(listD)
df['col4'] = df['col3'] + df['col1'].astype(str)
df

** ④ Change line name ** `df.index = ['A','B',,,]`

`Rename line`


df.index=['row0','row1','row2','row3','row4','row5']
df

This completes.

## 4. Create and read as a csv file

** ▼ Procedure ** ① Create a table with a csv file ② Import with read_csv method └ Specify heading column

** ① Create a table with a csv file ** Create a table like the one below.

** ② Import with read_csv method ** `read_csv ('falpath', index_col = n)` └ "'File path'": Absolute path or relative path is OK └ "index_col = n": Specify the column number (n: integer) to be the heading
** ▼ When reading the test.csv file on the desktop **

`Read csv file`


import pandas as pd
pd.read_csv('~/desktop/test.csv', index_col=0)

This completes.

Click here for a list of options for reading csv files (https://qiita.com/yuta-38/items/e1e890a647e77c7ccaad)

## 5. List of methods used ・ `Pd.DataFrame (list)` └ Convert list to table

・ Pd.DataFrame ([A, B ,,,]) └ Convert one-dimensional list to table at once

・ Df.index = ['A','B' ,,,] └ Change line name

・ Df.columns = ['a','b' ,,,] └ Change column name

・ Pd.DataFrame (list, index = ['A','B' ,,,]) └ Specify the row name when creating the table

・ Pd.DataFrame (list, columns = ['a','b' ,,,]) └ Specify the column name when creating the table

・ pd.DataFrame ({'A': [1,2 ,,],'B': [3,4 ,,],}) └ Specify the column name when creating the table └ Describe 'column name': [array] in ({})

・ Set_index ('a') └ Specify the column that will be the row name

・ Df ['b'] = pd.DataFrame (list) └ Add column (column name b) to table (df)

・ Df.T └ Transpose

・ Df ['B'] = df ['A'] / 2 └ "B": Column name to add └ "df ['A']" / 2: Substitute the value of the existing column name A divided by 2.

・ Df ['C'] = df ['A'] + df ['B'] └ Combine the values in columns "A" and "B" └ * str + int is an error (match the type)

・ Df ['A']. Astype (str) └ Change column "A" to "str" type

・ [['A','B' ,,], ['C','D' ,,] ,,,] └ Two-dimensional array

[Return to top](#)