1. DataFrame 处理缺失值 dropna()
df2.dropna(axis=0, how='any', subset=[u'ToC'], inplace=True)
把在ToC列有缺失值的行去掉
补充：还可以用df.fillna()来把缺失值替换为某个特殊标记
df = df.fillna("missing") # 用字符串替代 df = df.fillna(df.mean()) # 用均值或者其它描述性统计值替代
2. 根据某维度计算重复的行 duplicated()、value_counts()
print df.duplicated(['name']).value_counts() # 如果不指定列，默认会判断所有列 """ 输出: False 11118 True 664 表示有664行是重复的 """
duplicated()方法返回一个布尔型的Series，显示各行是否为重复行，非重复行显示为False，重复行显示为True
value_counts()方法统计数组或序列所有元素出现次数，对某一列统计可以直接用df.column_name.value_counts()
3. 去重 drop_duplicates()
df.drop_duplicates(['name'], keep='last', inplace=True) """ keep : {‘first’, ‘last’, False}, default ‘first’ first : Drop duplicates except for the first occurrence. last : Drop duplicates except for the last occurrence. False : Drop all duplicates. """
4. 拼接
(1) 拼接列 merge()
result = pd.merge(left, right, on='name', how='inner') """ 其它参数： pd.merge(left, right, how='inner', on=None, left_on=None, right_on=None, left_index=False, right_index=False, sort=False, suffixes=('_x', '_y'), copy=True, indicator=False) Examples >>> A >>> B lkey value rkey value 0 foo 1 0 foo 5 1 bar 2 1 bar 6 2 baz 3 2 qux 7 3 foo 4 3 bar 8 >>> A.merge(B, left_on='lkey', right_on='rkey', how='outer') lkey value_x rkey value_y 0 foo 1 foo 5 1 foo 4 foo 5 2 bar 2 bar 6 3 bar 2 bar 8 4 baz 3 NaN NaN 5 NaN NaN qux 7 """
其它参考：Merge, join, and concatenate
(2) 拼接行
def concat_by_row(data_dir, fout): dfs = [] for filename in os.listdir(data_dir): dfs.append(pd.read_excel(os.path.join(data_dir, filename))) print(dfs[-1].shape, filename) df = pd.concat(dfs, axis=0, ignore_index=True) # axis=0按行拼接；axis=1按列拼接 print(df.shape) df.to_excel(fout, index=False)
5. 找出在某一特定维度为空值的所有行
bool_arr = df.name.notnull() print bool_arr.value_counts() for idx, value in bool_arr.iteritems(): if not value: print ' ', idx, value print df.iloc[idx]
6. 指定dataframe的维度及顺序； 保存数据csv文件
res = {'name':[], 'buss':[], 'label':[]}
with codecs.open(fname, encoding='utf8') as fr:
for idx, line in enumerate(fr):
item = json.loads(line)
res['name'].append(item['name'])
res['buss'].append(item['buss'])
res['label'].append(item['label'])
df = pd.DataFrame(res, columns=['name', 'buss', 'label'])
df.to_csv('data/xxx.csv', index=False, encoding='utf-8')
7. 保存到文件
7.1 读写excel/csv格式文件
import pandas as pd def dataframe_read_and_write(fin, fout): # 读取fin文件，添加一列"新应答" df = pd.read_excel(fin) # df = read_csv(fin, encoding='utf-8') print df.head() fields = [u"序号", u"问题描述", u"原始应答", u"新应答"] data_out = defaultdict(list) for idx, row in df.iterrows(): try: row = row.to_dict() new_answer = "xxxxxx" for field in fields[:-1]: data_out[field].append(row[field]) data_out[fields[-1]].append(new_answer) except Exception as error: print "Error line", idx, error df_out = pd.DataFrame(data_out, columns=fields) df_out.to_excel(fout, sheet_name="Sheet1", index=False, header=True) # df_out.to_csv(fout, index=False, encoding="utf-8") if __name__ == '__main__': dataframe_read_and_write(fin="data/tmp.xlsx", fout="data/tmp_out.xlsx")
7.2 将多张DataFrame表写入到同一个excel文件的不同sheet中
import pandas as pd writer = pd.ExcelWriter('foo.xlsx') df.to_excel(writer, 'Data 0') df.to_excel(writer, 'Data 1') writer.save()
7.3 向一个已经存在的excel文件中写入一张新sheet；如果文件不存在则创建一个新文件再写入
import pandas from openpyxl import load_workbook def add_new_sheet(df, fout, sheet_name='Sheet1', columns=None): if fout and os.path.exists(fout): book = load_workbook(fout) writer = pd.ExcelWriter(fout, engine='openpyxl') writer.book = book writer.sheets = dict((ws.title, ws) for ws in book.worksheets) else: writer = pd.ExcelWriter(fout) df.to_excel(writer, sheet_name=sheet_name, columns=columns, index=False) writer.save() add_new_sheet(df, fout='Masterfile.xlsx', sheet_name="Main", columns=['Diff1', 'Diff2'])
参考：官方解决方案https://github.com/pandas-dev/pandas/issues/3441
7.4 读取excel文本中的多个sheet
import xlrd workbook = xlrd.open_workbook(fin) for sheet in workbook.sheets(): df = pd.read_excel(fin, sheet_name=sheet.name, index_col=None)
8. 排序
def sort_dataframe(df, fields_to_sort, fout=None): df = df.sort_values(by=fields_to_sort, ascending=True) if fout: df.to_excel(fout, index=False) return df df = pd.read_excel(data_file) sort_dataframe(df, fields_to_sort=["column_A", "column_B"], fout=data_file) df = pd.read_excel(data_file) # note: index改变，需要从文件重新读取，才会是有序的，后面遍历df的时候才不会出问题 print(df.head(10))
9. 轴标签重命名 df.rename()（列重命名、行index重命名）
>>> df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6]})
>>> df.rename(index=str, columns={"A": "a", "B": "c"})
a c
0 1 4
1 2 5
2 3 6
>>> df.rename(index=str, columns={"A": "a", "C": "c"})
a B
0 1 4
1 2 5
2 3 6
Using axis-style parameters
>>> df.rename(str.lower, axis='columns')
a b
0 1 4
1 2 5
2 3 6
>>> df.rename({1: 2, 2: 4}, axis='index')
A B
0 1 4
2 2 5
4 3 6
参数说明：
Signature: df.rename(mapper=None, index=None, columns=None, axis=None, copy=True, inplace=False, level=None) Docstring: Alter axes labels. Function / dict values must be unique (1-to-1). Labels not contained in a dict / Series will be left as-is. Extra labels listed don't throw an error. See the :ref:`user guide <basics.rename>` for more. Parameters ---------- mapper, index, columns : dict-like or function, optional dict-like or functions transformations to apply to that axis' values. Use either ``mapper`` and ``axis`` to specify the axis to target with ``mapper``, or ``index`` and ``columns``. axis : int or str, optional Axis to target with ``mapper``. Can be either the axis name ('index', 'columns') or number (0, 1). The default is 'index'. copy : boolean, default True Also copy underlying data inplace : boolean, default False Whether to return a new DataFrame. If True then value of copy is ignored. level : int or level name, default None In case of a MultiIndex, only rename labels in the specified level. Returns ------- renamed : DataFrame See Also -------- pandas.DataFrame.rename_axis
10. 数据选取，修改，切片
10.1 loc
在知道列名字的情况下，df.loc[index,column] 选取指定行，列的数据
# df.loc[index, column_name],选取指定行和列的数据 df.loc[0,'name'] # 'Snow' df.loc[0:2, ['name','age']] #选取第0行到第2行，name列和age列的数据, 注意这里的行选取是包含下标的。 df.loc[[2,3],['name','age']] #选取指定的第2行和第3行，name和age列的数据 df.loc[df['gender']=='M','name'] #选取gender列是M，name列的数据 df.loc[df['gender']=='M',['name','age']] #选取gender列是M，name和age列的数据
10.2 iloc
在column name特别长或者index是时间序列等各种不方便输入的情况下，可以用iloc (i = index), iloc完全用数字来定位 iloc[row_index, column_index]
df.iloc[0,0] #第0行第0列的数据，'Snow' df.iloc[1,2] #第1行第2列的数据，32 df.iloc[[1,3],0:2] #第1行和第3行，从第0列到第2列（不包含第2列）的数据 df.iloc[1:3,[1,2] #第1行到第3行（不包含第3行），第1列和第2列的数据
更多参考：
https://blog.csdn.net/yoonhee/article/details/76168253
11. 判断某个cell是否为空
if str(line["col_a"]).strip() == "nan": pass
12. Dataframe值替换
df["col_a"] = df["col_a"].replace({"b": "C", "e": "G"})
更多参考：https://jingyan.baidu.com/article/454316ab4d0e64f7a6c03a41.html
13. Dataframe筛选数据
df2 = df[df["col_a"] == "cc"] # 等于某个值 df3 = df[df["col_a"].isin(["bb", "cc", "ee"])] # 取值在某个范围内
更多参考：https://jingyan.baidu.com/article/0eb457e508b6d303f0a90572.html
14. 其它常用操作
# df = pd.read_csv("../../data/data_part1.txt", sep="$") df = pd.read_csv("data/data_part1.csv", low_memory=False) # 数据概览 df.info() df.describe() # ==> 只显示float型维度的[count, mean, std, min]等统计信息, 例如0108, 3816, 2453, 0112, 2428, 2304 # 数据查看 df.head(n=5) # 查看开头几行, 默认n=5 df.tail(n=5) # 查看末尾几行, 默认n=5 df.shape # 查看行列维度 df.columns # 查看列名和列数 df.dtypes # 查看数据类型 ==> 可以看到哪些维度的数值是object型/float型 df["0108"].hist() # 查看变量分布 df["0108"].unique() # 查看有哪些取值 df["0108"].value_counts() # 查看这一列的值统计 # 缺失值统计 df.isnull().sum() # 查看每一列缺失值情况 df["n_null"] = df.isnull().sum(axis=1) # 查看每一行缺失值情况 # 缺失值填充 mode_df = df.fillna(df.mode().iloc[0], inplace=True) # 用众数填充 median_df = df.fillna(df.median()) # 用中位数填充 df["0108"][df.vid.isnull()] = "0" # 对某一列填充
15. 遇到的问题和解决方法
15.1 df.to_excel(fout) 报错"openpyxl.utils.exceptions.IllegalCharacterError"
(step 1) pip install xlsxwriter
(step 2) df.to_excel(fout, engine="xlsxwriter")
15.2 保存文件时报错"UserWarning: Ignoring URL 'http://www.xxxxxxx' with link or location/anchor > 255 characters since it exceeds Excel's limit for URLS"
writer = pd.ExcelWriter(fout, engine="xlsxwriter", options={'strings_to_urls': False}) df.to_excel(writer, index=False) writer.save()
参考：