How to keep updating new values in pandas data frame

I am trying to check the price change of all the crypto at a 15-minute interval. Till now, I have been able to fetch all coins traded with USDT pair, get their price change percentage, and sort then descending order. Now the plan is to wait for 15 min, again fetch the data, and compare it with the old one. This will keep on happening every 15 minutes. (For example, let’s say I fetched data for 5.30, at 5.45 again data will be fetched – comparison will be made and the most volatile coin will be chosen. At 6.00, the data will be fetched again, now it has to be compared with data of 5.45 and so on.)

This is what I have till now (This is based on CCXT library for python)This code works for one-time fetch, all I need is some guidance on how to do comparison and store/update new values.Should I create new file pricechange2 and keep comparing with it or should I keep on adding and deleting columns in dataframe?

markets=exchange.fetchTickers()
with open('pricechange1.txt', 'w') as json_file:
    json.dump(markets, json_file)
with open('pricechange1.txt') as json_file:
    data = json.load(json_file)

symbol_list = [value['symbol'] for value in data.values()]

r = re.compile(".*/USDT")
filtered_list = list(filter(r.match, symbol_list))
fl = list([x for x in filtered_list if "DOWN" not in x and "UP" not in x and "BULL" not in x and "BEAR" not in x])
pp1 = []
for i in fl:
    pp = data[i]['info']['priceChangePercent']
    pp1.append(float(pp))
df = pd.DataFrame({'coinpair': fl})
df['change1'] = pp1
df = df.sort_values(by='change1', ascending=False)
print(df['coinpair'].head(10).iloc[0])
print(df.head(10))

Current output of above code

      coinpair  change1
90    FTT/USDT   29.677
48    FTM/USDT   24.615
246  TORN/USDT   24.508
259  MINA/USDT   23.265
156  YFII/USDT   20.739
37   CELR/USDT   18.435
179   INJ/USDT   17.601
207  FIRO/USDT   15.848
206   TWT/USDT   15.715
126  VTHO/USDT   15.634

Answer

I would recommend putting those things into separate functions and using schedule library to run them periodically (in your case in every 15 mins). First of all, you need to install it using:

pip install schedule

Then import it in your script:

import schedule
import time
...

Now, put everything into a separate function:

markets = None
def fetch_data():
    global markets
    markets =  exchange.fetchTickers()
    return markets

def write_to_file(data):
    with open('pricechange1.txt', 'w') as json_file:
        json.dump(data, json_file)

previous_pp = None
def form_df(data, previous_pp):
    global previous_pp
    symbol_list = [value['symbol'] for value in data.values()]

    r = re.compile(".*/USDT")
    filtered_list = list(filter(r.match, symbol_list))
    fl = list([x for x in filtered_list if "DOWN" not in x and "UP" not in x and "BULL" not in x and "BEAR" not in x])
    pp1 = []
    
    for i in fl:
        if previous_pp:
            current_pp = data[i]['info']['priceChangePercent']
            if previous_pp > current_pp:
                # do something
                # update previous_pp value
            else:
                # do something
                # update previous_pp value
        else:
            previous_pp = current_pp = data[i]['info']['priceChangePercent']
            pp1.append(float(current_pp))

    df = pd.DataFrame({'coinpair': fl})
    df['change1'] = pp1
    df = df.sort_values(by='change1', ascending=False)
    print(df['coinpair'].head(10).iloc[0])
    print(df.head(10))

And run them periodically as below:

schedule.every(15).minutes.do(fetch_data)
schedule.every(15).minutes.do(write_to_file, data=markets)
schedule.every(15).minutes.do(form_df, data=markets, previous_p=previous_pp)

while True:
    schedule.run_pending()
    time.sleep(1)