Hi Carl,

I had played around with the HBase thrift api and it was a matter of generating 
the cpp files and using them.
I assumed Hive thrift would be something similar - hence my messing around with 
that.

But you correctly guessed my intentions : a c++ client. So i'll concentrate on 
hiveclient.

Thanks
J

From: Carl Steinbach <[email protected]>
Date: November 30, 2010 9:00:16 PM PST
To: [email protected]
Subject: Re: Deserializing results from ThriftHiveClient::fetch*
Reply-To: [email protected]



> Hi Joy,
> 
> I'm not sure what you're trying to accomplish (maybe you can provide some 
> more details?),
> but assuming that your goal is to access Hive via a C++ Thrift client, then I 
> strongly recommend
> that you use the hiveclient wrapper code located in odbc/src/cpp instead of 
> messing around
> with the raw code generated by the Thrift compiler. Documentation for the C++ 
> Hive client library
> is located in the header file (odbc/src/cpp/hiveclient.h).
> 
> However, if you really do want to access the C++ Thrift API directly, then 
> you should look at the
> implementation of the hiveclient library for examples of how to unmarshall 
> the results returned by
> fetchOne().
> 
> Thanks.
> 
> Carl
> 
> On Tue, Nov 30, 2010 at 4:51 PM, Saptarshi Guha 
> <[email protected]> wrote:
> Hello
> 
> A few questions. Once i get the result from e.g   void fetchOne(std::string& 
> _return);
> how does one parse the bytes in _return?
> 
> Is this what getSchema is for? Do I call getSchema, figure out the column 
> types
> and (somehow) deserialize? Or are there functions for to deserialize this?
> The source does say something about Thrift serialized form, so maybe there 
> could
> thrift functions to deserialize the return values from fetch*.
> 
> Once again thanks for your pointers
> 
> Cheers
> Joy
> 
> 

Reply via email to