使用Ibis python将数据插入到impala表中

Question

我正在尝试将df插入到ibis创建的带分区的impala表中。我在远程内核上运行，在Windows10机器上使用spyder 3.2.4，在运行CentOS的边缘节点机器上使用Python3.6.2。

我得到以下错误：

Writing DataFrame to temporary file
Writing CSV to: /tmp/ibis/pandas_0032f9dd1916426da62c8b4d8f4dfb92/0.csv
Traceback (most recent call last):

File "/usr/local/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2910, in run_code
exec(code_obj, self.user_global_ns, self.user_ns)

File "", line 1, in 
insert = target_table.insert(df3)

File "/usr/local/lib/python3.6/site-packages/ibis/impala/client.py", line 1674, in insert
writer, expr = write_temp_dataframe(self._client, obj)

File "/usr/local/lib/python3.6/site-packages/ibis/impala/pandas_interop.py", line 225, in write_temp_dataframe
return writer, writer.delimited_table(path)

File "/usr/local/lib/python3.6/site-packages/ibis/impala/pandas_interop.py", line 188, in delimited_table
schema = self.get_schema()

File "/usr/local/lib/python3.6/site-packages/ibis/impala/pandas_interop.py", line 184, in get_schema
return pandas_to_ibis_schema(self.df)

File "/usr/local/lib/python3.6/site-packages/ibis/impala/pandas_interop.py", line 219, in pandas_to_ibis_schema
return schema(pairs)

File "/usr/local/lib/python3.6/site-packages/ibis/expr/api.py", line 105, in schema
return Schema.from_tuples(pairs)

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 109, in from_tuples
return Schema(names, types)

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 55, in init
self.types = [validate_type(typ) for typ in types]

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 55, in 
self.types = [validate_type(typ) for typ in types]

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 1040, in validate_type
return TypeParser(t).parse()

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 901, in parse
t = self.type()

File "/usr/local/lib/python3.6/site-packages/ibis/expr/datatypes.py", line 1033, in type
raise SyntaxError('Type cannot be parsed: {}'.format(self.text))

File "", line unknown
SyntaxError: Type cannot be parsed: integer

Answer 1

由于hadoop系统的结构和安全性的原因，出现了错误。Ibis包尝试在__ibis_tmp & /tmp/ibis/中分别创建temp_db和temp_hdfs_location。因为在我们的系统中，除了root/system admin之外，任何用户都不能访问默认位置...从/tmp/ibis/获取数据到实际数据库时，insert命令出错(仍不清楚，但可能是通过__ibis_tmp数据库)。一旦我们将ibis包的config_init.py文件编辑到一个允许的临时位置/db。它就像一个护身符一样有效。

Answer 2

而不是编辑前面提到的config_init.py

https://stackoverflow.com/a/47543691/5485370

使用ibis.options可以更轻松地分配临时数据库和路径：

ibis.options.impala.temp_db = 'your_temp_db'
ibis.options.impala.temp_hdfs_path = 'your_temp_hdfs_path'