[英]How do I convert a .dbf file into a Pandas DataFrame?
我有一个.dbf
文件,我想将其转换为 pandas DataFrame
,但DataFrame
无法直接转换数据。
使用我的dbf
library ,以下 function 将完成这项工作:
def dbf_to_dataframe(filename):
"""
converts the dbf table at filename into a Panda's DataFrame
data types and field names are preserved
"""
import dbf
import numpy as np
import pandas as pd
from datetime import date, datetime, time
names = []
types = []
table = dbf.Table(filename)
for name in table.field_names:
ftype, size, decimals, _ = table.field_info(name)
ftype = chr(ftype)
if ftype in 'GP':
continue
if ftype == 'N' and decimals:
ftype = 'F'
dtype = {
'B': 'float64',
'C': 'string',
'D': 'datetime64[ns]',
'F': 'float64',
'I': 'int64',
'L': 'boolean',
'M': 'string',
'N': 'int64',
'T': 'datetime64[ns]',
'Y': 'float64',
}[ftype]
names.append(name)
types.append(dtype)
with table:
series = [[] for _ in names]
for rec in table:
for i, value in enumerate(rec):
if isinstance(value, date):
value = datetime.combine(value, time())
elif value is None:
value = np.nan
series[i].append(value)
data_recs = dict(
(n, pd.Series(s, dtype=t))
for n, s, t in zip(names, series, types)
)
return pd.DataFrame(data_recs)
我真的很欣赏上面答案中的代码,有人可以解释为什么要这样组装代码吗? 我查找了 dbf 模块文档,但没有找到。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.