여러 데이터 쉽게 불러오기

여러 데이터 쉽게 불러오기.ipynb

0.01MB

import pandas as pd
from time import time

# Load data (pandas version)
start = time()
train = pd.read_csv('./data/transactions_train.csv')
end = time()
print(f"데이터 로드 시간 : {end-start}초")
train

# Check memory usage
mem_usage = train.memory_usage(deep=True).sum() / 1024 / 1024 / 1024
print(f"Memory Usage : {mem_usage:.4} GiB")

# parquet로 다시 만들어봅시다. 데이터의 일부만 가져옵니다.
part = pd.read_csv('./Downloads/transactions_train.csv',
                  nrows = 1000)
part2 = pd.read_csv('./Downloads/transactions_train.csv',
                  usecols=['t_dat','sales_channel_id'])
part2

sales = part["sales_channel_id"].value_count() * 0

for chunk in pd.read_csv('./Downloads/transactions_train.csv',
                        chunksize=3000000):
    print(chunk["sales_channel_id"].value_counts())
    sales = sales + chunk["sales_channel_id"].value_counts()
sales

train.loc[train["t_dat"] > '2020-06-01']