采购实用python数据处理方法

正则表达式

匹配多位数字

re.findall(r'\d{n}', str)
# 返回 n 位数字

匹配小数

re.findall(r'\d.+', str)
# 返回小数

list

判断列表是否为空

list == []

len(list) = 0 # ==> []
len(list) > 1 # ==> 非空

$$axis=\{\begin{array}{ll}0,& \text{列}\\ 1,& \text{行}\end{array}$$

从后向前取n个值

arr[-n:]

获取 value 最大元素的 index

aa = [...]
aa.index(max(aa))

迭代函数

enumerate(可迭代对象)
# 生成 index 和 value

pandas

合并

pd.concat([data, pd.DataFrame(columns = blank)], axis = 0) # 增加空白columns

删除指定column为空的行

data.dropna(subset = ['price'])

删除某一列是重复的

df.drop_duplicates(subset = 'label')

判断元素是否为空

pd.isna()
pd.isnull
pd.notnull()

将某一列为空的值取出来

df[pd.isnull(df)]

replace 替换多个

df.replace({'A':100, 'B':50})

排序

df.sort_values(by = ['a'])

删除空白列

df.dropna(axis = 1, how = 'all')
# axis = 0 按行(default)
# axis = 1 按列

# any : 只要有 1 个
# all : 全部都为空才 drop

取最大/最小值的 index

df.idxmax() # 最大值的索引
df.idxmin() # 最小值的索引

使用plotly画 dataframe

import plotly.express as px

px.scatter(data, x = '', y = '') # x 和 y 是 data 中的 columns 

数据类型转换

判断数据类型

isinstance(a, str) # 判断 a 是否为 str型

字符串 ==> 数值

int(str)

ndarry 转 list

aa = np.array([])
bb = aa.tolist()

列标转 Excel 列标字母（调用 xlsxwriter）

xlsxwriter.utility.xl_col_to_name(index)

参考资料：Convert spreadsheet number to column letter

• xlsxwriter 源码方法：

def xl_col_to_name(col, col_abs=False):
    """
    Convert a zero indexed column cell reference to a string.

    Args:
       col:     The cell column. Int.
       col_abs: Optional flag to make the column absolute. Bool.

    Returns:
        Column style string.

    """
    col_num = col
    if col_num < 0:
        warn("Col number %d must be >= 0" % col_num)
        return None

    col_num += 1  # Change to 1-index.
    col_str = ''
    col_abs = '$' if col_abs else ''

    while col_num:
        # Set remainder from 1 .. 26
        remainder = col_num % 26

        if remainder == 0:
            remainder = 26

        # Convert the remainder to a character.
        col_letter = chr(ord('A') + remainder - 1)

        # Accumulate the column letters, right to left.
        col_str = col_letter + col_str

        # Get the next order of magnitude.
        col_num = int((col_num - 1) / 26)

    return col_abs + col_str

• stackoverflow 方法

start_index = 1   #  it can start either at 0 or at 1
letter = ''
while column_int > 25 + start_index:   
    letter += chr(65 + int((column_int-start_index)/26) - 1)
    column_int = column_int - (int((column_int-start_index)/26))*26
letter += chr(65 - start_index + (int(column_int)))