How to sort dataframe result based on column values

Question

import requests
import pandas as pd

url = "https://coinmarketcap.com/new/"
page = requests.get(url,headers={'User-Agent': 'Mozilla/5.0'}, timeout=1)
pagedata = page.text
usecols = ["Name", "Price", "1h", "24h", "MarketCap", "Volume"]#, "Blockchain"]

df = pd.read_html(pagedata)[0] #Checking table
df[["Name", "Symbol"]] = df["Name"].str.split(r"\d+", expand=True)
df = df.rename(columns={"Fully Diluted Market Cap": "MarketCap"})[usecols]
dfAsString = df.to_string(index=False)

print(dfAsString)

Current Code Output: (Truncated)

               Name            Price      1h      24h       MarketCap       Volume
0        DollarPepe         $0.02752  22.64%  336.25%              $3     $456,913
1       Billy Token      $0.00002822  41.69%   75.80%      $1,958,942   $6,999,241
2              JEFF          $0.1946   4.42%  226.18%     $19,458,328  $19,744,583
3            PUG AI   $0.00000001459  10.80%   15.84%      $1,459,428     $239,454
4         FART COIN    $0.0000004281   1.13%   42.13%     $42,806,075      $46,604
[30 rows x 6 columns] 

How to produce the outpur sorted based on a particular column (24h)? -> Truncated

               Name            Price      1h      24h       MarketCap       Volume
0        DollarPepe         $0.02752  22.64%  336.25%              $3     $456,913
2              JEFF          $0.1946   4.42%  226.18%     $19,458,328  $19,744,583
1       Billy Token      $0.00002822  41.69%   75.80%      $1,958,942   $6,999,241
4         FART COIN    $0.0000004281   1.13%   42.13%     $42,806,075      $46,604
3            PUG AI   $0.00000001459  10.80%   15.84%      $1,459,428     $239,454
[30 rows x 6 columns]

Answer 1

I would convert all the numeric columns in your df to numeric values; then you can sort them easily (you can always add back $ and % as required on display).

numcols = df.columns[df.columns != 'Name']
df[numcols] = df[numcols].apply(lambda c:pd.to_numeric(c.str.replace(r'[^\d.]|(?<!\d)\.|\.(?!\d)', '', regex=True)))
df = df.sort_values('24h', ascending=False)

Output (for your sample data):

          Name         Price     1h     24h  MarketCap    Volume
0   DollarPepe  2.752000e-02  22.64  336.25          3    456913
2         JEFF  1.946000e-01   4.42  226.18   19458328  19744583
1  Billy Token  2.822000e-05  41.69   75.80    1958942   6999241
4    FART COIN  4.281000e-07   1.13   42.13   42806075     46604
3       PUG AI  1.459000e-08  10.80   15.84    1459428    239454

Note the non-numeric character replacement is more complicated than just [^\d.] as would be implied by your sample data; this is because some of the other price values fetched from that page have ... in them (presumably because they are too small to represent). Ideally you need to figure out how to get those as exact values; otherwise they can only be approximated.

Alternatively, you can sort on the values by converting them to floats before passing the series to sort_values:

df = df.sort_values('24h', ascending=False, key=lambda v:v.str.replace('%', '').astype(float))

Output:

          Name           Price      1h      24h    MarketCap       Volume
0   DollarPepe        $0.02752  22.64%  336.25%           $3     $456,913
2         JEFF         $0.1946   4.42%  226.18%  $19,458,328  $19,744,583
1  Billy Token     $0.00002822  41.69%   75.80%   $1,958,942   $6,999,241
4    FART COIN   $0.0000004281   1.13%   42.13%  $42,806,075      $46,604
3       PUG AI  $0.00000001459  10.80%   15.84%   $1,459,428     $239,454