python连接impala报错

背景

连接器一般使用impyla，但是我发现这个库问题不少，关键报错了看不出来哪里出问题了，着实令人蛋疼！！

问题

常见的难以解决的报错有：
1. AttributeError: 'TBufferedTransport' object has no attribute 'trans'
2. TypeError: expecting list of size 2 for struct args
3. 见https://github.com/cloudera/impyla/issues/286

解决办法

这实际上都是同一个问题，pip install thrift==0.9.3即可

python3时的报错

如果你的python版本是3的话，会发现报错可能不一样，比如下面的：

Traceback (most recent call last):
  File "impala_py.py", line 68, in <module>
    df = run_sql(sql)
  File "impala_py.py", line 29, in run_sql
    conn = connect(host='xxx', auth_mechanism='PLAIN', port=21050, user='xxx', password='xxxx')
  File "/home/liming/anaconda2/envs/test3/lib/python3.6/site-packages/impala/dbapi.py", line 147, in connect
    auth_mechanism=auth_mechanism)
  File "/home/liming/anaconda2/envs/test3/lib/python3.6/site-packages/impala/hiveserver2.py", line 758, in connect
    transport.open()
  File "/home/liming/anaconda2/envs/test3/lib/python3.6/site-packages/thrift_sasl/__init__.py", line 67, in open
    if not self._trans.isOpen():
AttributeError: 'TSocket' object has no attribute 'isOpen'

目前尝试可以的解决办法是安装如下版本的包：
thrift-sasl==0.2.1
只在linux尝试成功了，windows下安装失败