is there any way in Python to change the columns' positions?

Question

enter image description here

Hi, is there any way in Python for me to change the columns' positions like below: enter image description here

Thanks a lot!

Update: I have been trying to reframe the columns in the dump way, any code I can apply to increase the efficiency? Thanks.

now = dt.datetime.now()
length = relativedelta(weeks = 8)
start = now - length
end = now
stocks = ['TSLA','AMZN','FB']
tokens = ['BTC-USD','ETH-USD','DOGE-USD','XRP-USD']

#2. Get data and arrange them together without NaNs
stocks = web.DataReader(stocks, 'yahoo', start, end)
tokens = web.DataReader(tokens, 'yahoo', start, end)
df = tokens.merge(stocks, how='inner', right_index = True, left_index = True)
df.dropna(inplace = True)
df = df[[('Open', 'TSLA'),('High', 'TSLA'),('Low', 'TSLA'),('Close', 'TSLA'),('Adj Close', 'TSLA'),('Volume', 'TSLA'),
                          ('Open', 'AMZN'),('High', 'AMZN'),('Low', 'AMZN'),('Close', 'AMZN'),('Adj Close', 'AMZN'),('Volume', 'AMZN'),
                          ('Open', 'FB'),('High', 'FB'),('Low', 'FB'),('Close', 'FB'),('Adj Close', 'FB'),('Volume', 'FB'),
                          ('Open', 'BTC-USD'),('High', 'BTC-USD'),('Low', 'BTC-USD'),('Close', 'BTC-USD'),('Adj Close', 'BTC-USD'),('Volume', 'BTC-USD'),
                          ('Open', 'ETH-USD'),('High', 'ETH-USD'),('Low', 'ETH-USD'),('Close', 'ETH-USD'),('Adj Close', 'ETH-USD'),('Volume', 'ETH-USD'),
                          ('Open', 'DOGE-USD'),('High', 'DOGE-USD'),('Low', 'DOGE-USD'),('Close', 'DOGE-USD'),('Adj Close', 'DOGE-USD'),('Volume', 'DOGE-USD'),
                          ('Open', 'XRP-USD'),('High', 'XRP-USD'),('Low', 'XRP-USD'),('Close', 'XRP-USD'),('Adj Close', 'XRP-USD'),('Volume', 'XRP-USD')]]
df.to_csv('Data/newohlc.csv')

you have used the pandas and dataframe tags but the screenshots you are showing are of excel. i assume you want to change your df column order which you are then outputting as a csv? have you checked this thread? https://stackoverflow.com/questions/13148429/how-to-change-the-order-of-dataframe-columns#:~:text=You%20need%20to%20create%20a,columns%20in%20this%20new%20order. — Scinana, May 25 '21 at 11:19

score 0 · Answer 1 · answered May 25 '21 at 14:56

You can use the syntax DataFrame[["column1", "column2", "column3"]] with the column names in the desired order to reorder the columns. Example:

import pandas as pd
import numpy as np
df1 = {
    'Name':['George','Andrea','micheal','maggie','Ravi',
               'Xien','Jalpa'],
     'Gender':["M","F","M","F","M","M","F"],      
    'Score':[62.7,47.7,55.6,74.6,31.5,77.3,85.4],
    'Rounded_score':[63,48,56,75,32,77,85]
   }
df1 = pd.DataFrame(df1,columns=['Name','Gender','Score','Rounded_score'])
print(df1)
df2=df1.reindex(columns= ['Rounded_score', 'Gender', 'Score','Name'])
print(df2)

thanks, but if there are many columns, any way to reframe them with a higher efficiency? — Ka Kin Li, May 25 '21 at 15:54

norie · Answer 2 · 2021-05-25T18:52:06.800

If you want to generate the list of tuples you are using to position/rename the columns you could use something like this.

Note, there's probably a better way than using a double loop and I did have something working using zip but when using that method I wasn't able to get the sorting right..

For example, all the open prices were grouped together rather than all the prices etc. being grouped by stock/token.


import datetime as dt
from dateutil.relativedelta import *
import pandas_datareader as web

now = dt.datetime.now()
length = relativedelta(weeks = 8)
start = now - length
end = now
stocks = ['TSLA','AMZN','FB']
tokens = ['BTC-USD','ETH-USD','DOGE-USD','XRP-USD']
cols = ['Open','High', 'Low', 'Close', 'Adj Close' , 'Volume']

newcols = []

for symbol in stocks+tokens:
  for col in cols:
    newcols.append((col, symbol))

print(newcols)

#2. Get data and arrange them together without NaNs
stock_data = web.DataReader(stocks, 'yahoo', start, end)
token_data = web.DataReader(tokens, 'yahoo', start, end)
df = token_data.merge(stock_data, how='inner', right_index = True, left_index = True)

print(df.head())

df.dropna(inplace = True)

df = df[newcols]

print(df.head())

I've found an alternative to using the double loop but I'm no sure it's 100% right.

I'll post it any way - need some fresh eyes to check it.


newcols = list(tuple(zip(cols*len(stocks+tokens), (stocks+tokens)*len(cols))))

newcols.sort(key = lambda x: (stocks+tokens).index(x[1])*10 + cols.index(x[0]))

is there any way in Python to change the columns' positions?

2 Answers2