系统:Windows 10 语言版本:conda 4.4.10 编辑器:JetBrains PyCharm Community Edition 2018.2.2 x64 pandas:0.22.0
Part 1:场景描述
["time", "pos", "value1"]
df
Part 2:代码逻辑
Part 3:代码
import pandas as pd
import numpy as np
# 显示所有列
pd.set_option('display.max_columns', None)
# 显示所有行
pd.set_option('display.max_rows', None)
# 设置显示长度为100
pd.set_option('max_colwidth', 100)
# 设置对齐
pd.set_option('display.unicode.ambiguous_as_wide', True)
pd.set_option('display.unicode.east_asian_width', True)
# 设置打印宽度
pd.set_option('display.width', 180)
dict_1 = {"time": ["2019-11-02", "2019-11-15", "2019-11-25", "2019-11-05",
"2019-12-13", "2019-12-03", "2019-12-16", "2019-12-29"],
"pos": ["A", "A", "B", "B", "C", "C", "C", "D"],
"value1": [10, 20, 30, 40, 50, 60, 70, 80]}
df = pd.DataFrame(dict_1, columns=["time", "pos", "value1"])
print("\n", "df", "\n", df, "\n", df.dtypes)
df["time1"] = pd.to_datetime(df['time'])
print("\n", "df", "\n", df, "\n", df.dtypes)
# 获取年月日信息
df["年"] = df["time1"].dt.year
df["月"] = df["time1"].dt.month
df["日"] = df["time1"].dt.day
df["时"] = df["time1"].dt.hour
df["分"] = df["time1"].dt.minute
df["秒"] = df["time1"].dt.second
df["flag"] = df["日"]
df["xun"] = np.where((df["flag"] > 10) & (df["flag"] <= 20), "中旬", np.where(df["flag"] <= 10, "上旬", "下旬"))
print("\n")
print(df)
df_1 = df[df["xun"] == "上旬"]
print("\n")
print(df_1)
df_1 = df[df["xun"] == "中旬"]
print("\n")
print(df_1)
df_1 = df[df["xun"] == "下旬"]
print("\n")
print(df_1)
代码截图
Part 4:部分代码解读
df["time1"] = pd.to_datetime(df['time'])
时间格式转换,新生成的数据类型为datetime64
时间格式转换
df["日"] = df["time1"].dt.day
获取日期对应的具体几号df["xun"] = np.where((df["flag"] > 10) & (df["flag"] <= 20), "中旬", np.where(df["flag"] <= 10, "上旬", "下旬"))
,两重判断flag
列的每个元素进行计算,结果为xun
df_1 = df[df["xun"] == "上旬"]
获取上旬数据本文为原创作品,欢迎分享朋友圈
长按图片识别二维码,关注本公众号 Python 优雅 帅气