pandas中的解析函数：
| 函数名 | 说明 |
|---|---|
| read_csv() | 默认分隔符为逗号 |
| read_table() | 默认分隔符为制表符(‘/t’) |
| read_fwf() | 读取定宽列格式数据,也就是说没有分隔符 |
| read_clipboard() | 读取剪切板中的数据.在将网页转换为表格时很有用 |
path分隔符 sep 列名 header OR name 索引 index_col跳行 skiPRows缺失值 na_values行数 nrows文件块 chunksize还有很多,诸如comment, parse_dates, keep_date_col, converters, dayfirst, date_parser, iterator, skip_footer, verbose, encoding, squeeze, thousands
函数示范举例：
import pandas as pd
分隔符1>>> pd.read_table('ex3.txt',sep='/s+') A B Caaa -0.264438 -1.026059 -0.619500bbb 0.927272 0.302904 -0.032399ccc -0.264273 -0.386314 -0.217601ddd -0.871858 -0.348382 1.100491列名列名为None:
>>> pd.read_csv('ex2.csv',header=None) 0 1 2 3 40 1 2 3 4 hello1 5 6 7 8 world2 9 10 11 12 foo自定义列名:
>>> pd.read_csv('ex2.csv', names=['a','b','c','d','message']) a b c d message0 1 2 3 4 hello1 5 6 7 8 world2 9 10 11 12 foo索引注:如果希望列名中的某一个列做成DataFrame的索引,通过index_col参数指定.
>>> pd.read_csv('ex2.csv', names=['a','b','c','d','message'], index_col='message') a b c dmessage hello 1 2 3 4world 5 6 7 8foo 9 10 11 12当然我们也可以做一个层次化索引.
>>> pd.read_csv('csv_mindex.csv') key1 key2 value1 value20 one a 1 21 one b 3 42 one c 5 63 one d 7 84 two a 9 105 two b 11 126 two c 13 147 two d 15 16>>> pd.read_csv('csv_mindex.csv', index_col=['key1', 'key2']) value1 value2key1 key2 one a 1 2 b 3 4 c 5 6 d 7 8two a 9 10 b 11 12 c 13 14 d 15 16跳行2 >>> pd.read_csv('ex4.csv', skiprows=[0,2,3]) a b c d message0 1 2 3 4 hello1 5 6 7 8 world2 9 10 11 12 foo缺失值>>> res = pd.read_csv('ex5.csv')>>> res something a b c d message0 one 1 2 3 4 NaN1 two 5 6 NaN 8 world2 three 9 10 11 12 foo>>> pd.isnull(res) something a b c d message0 False False False False False True1 False False False True False False2 False False False False False False>>> pd.read_csv('ex5.csv', na_values=['NULL']) something a b c d message0 one 1 2 3 4 NaN1 two 5 6 NaN 8 world2 three 9 10 11 12 foo用字典为各列指定NA值
#可以看到[0][c],[2][message]变为NA值>>> dict={'c':[1000,3],'message':['NA','foo']}>>> pd.read_csv('ex5.csv', na_values=dict) something a b c d message0 one 1 2 NaN 4 NaN1 two 5 6 NaN 8 world2 three 9 10 11 12 NaN行数>>> pd.read_csv('ex6.csv', nrows=10) one two three four key0 0.467976 -0.038649 -0.295344 -1.824726 L1 -0.358893 1.404453 0.704965 -0.200638 B2 -0.501840 0.659254 -0.421691 -0.057688 G3 0.204886 1.074134 1.388361 -0.982404 R4 0.354628 -0.133116 0.283763 -0.837063 Q5 1.817480 0.742273 0.419395 -2.251035 Q6 -0.776764 0.935518 -0.332872 -1.875641 U7 -0.913135 1.530624 -0.572657 0.477252 K8 0.358480 -0.497572 -0.367016 0.507702 S9 -1.740877 -1.160417 -1.637830 2.172201 G文件块>>> chunk = pd.read_csv('ex6.csv', chunksize = 1000)>>> from pandas import Series>>> tot = Series([])#Series是一种类似于数组的对象，它由一组数据以及一组与之相关的数据标签（即索引）组成>>> for piece in chunk: #value_counts()统计这个块中相同值的个数 #块中缺失值几即赋值为0... tot = tot.add(piece['key'].value_counts(), fill_value=0)... >>> tot = tot.sort_values(ascending=False) >>> tot[:10]E 368X 364L 346O 343Q 340M 338J 337F 335K 334H 330dtype: float64可移步参考DataFrame或者Series.
新闻热点
疑难解答
