Question

使用python .csv's模块将其中一个pd.DataFrame转换为odo时，我会收到TypeError

    TypeError: Cannot cast array from dtype('float64') to dtype('int64') 
               according to the rule 'safe'

适用于其他csv's

的代码

# csv table file name
csvNm = 'table.csv'

# convert mysql table to csv
odo_csv = odo(tstConn.connect_string + '::' + tbl , csvNm)

# convert csv to pandas 
odo_df = odo(odo_csv , pd.DataFrame)

这是我到目前为止所做的尝试无效：

import pandas as pd
from odo import odo, resource, discover, convert

odo_csv=odo(tstConn.connect_string + '::' + tbl , csvNm)
csv=resource(csvNm)
ds=discover(csv)

# Convert csv to pandas
odo_df = odo(odo_csv , pd.DataFrame, dshape=ds)

和此：

odo_df = odo(odo_csv , pd.DataFrame, casting='unsafe')

更新1 看起来我忽略了这个错误中最明显的暗示

pandas\parser.pyx in pandas.parser.TextReader._convert_tokens (pandas\parser.c:11816)()

导致Windows SO中的编码问题。但这不是：

odo_df = odo(odo_csv , pd.DataFrame, encoding=odo_csv.encoding)

或这项工作

odo_df = odo(odo_csv , pd.DataFrame, encoding='cp1252')

这种不合时宜的方式（针对我的用例）来自pandas-reading-csv-files（与上面相同的链接）

# Python3
with open('/tmp/test.csv', 'r', encoding='cp1252') as f:
    df = pd.read_csv(f)
    print(df)

不确定下一步该尝试，我们将不胜感激。

Answer 1

有效的解决方案是：

import pandas as pd
from odo import odo, resource, discover, convert

# convert mysql to csv
odo_csv=odo(raw_dbConn.connect_string + '::' + tblName , csvNm, header=True)

# Get odo resource aka sqlalchemy.Table instance
resc=resource(raw_dbConn.connect_string + '::' + tblName )

# Discover the resc
ds=discover(resc)

# Convert csv to dataframe    
odo_df = odo(odo_csv , pd.DataFrame, dshape=ds ,encoding=odo_csv.encoding)

odo在csv和mysql之间转换数据

1 个答案: