hi Andrew, All the things you're looking for are in the arrow::ipc namespace
http://arrow.apache.org/docs/cpp/namespacearrow_1_1ipc.html The IO interfaces are in arrow::io http://arrow.apache.org/docs/cpp/namespacearrow_1_1io.html Your best bet would be to work through the C++ API for the Arrow object model (schemas, record batches, arrays) and look at the test suite for example usages. Here's an example patch I wrote recently for the MapD GPU database adding an Arrow RPC endpoint for loading data (this reads from a std::string, not an InputStream though): https://github.com/mapd/mapd-core/pull/93 I would love to have some user documentation for the C++ library; when the developer community grows a little larger I think we should have enough spare bandwidth to do this, but until now we have been a bit spread thin staying on top of JIRAs. - Wes On Mon, Oct 2, 2017 at 6:57 PM, Andrew Pham (BLOOMBERG/ 731 LEX) <apha...@bloomberg.net> wrote: > Hello everyone, my proof of concept successfully writes a bunch of bytes > representing a schema that is a list of objects (consisting of an integer and > a double, each) to a file on the system (like a .txt file). This is done in > Java, and I can also read from that file and interpret the data successfully > if need be. > > I'm wondering if anyone knows of any good examples in C++ on reading from a > WritableByteChannel (or something of the sort in C++), interpreting the data, > and printing it out (or even constructing C++ objects from it). That would > be much appreciated, thanks! > > Additionally, given the interpreted bytes, can we interpret the schema on the > C++ side (from the Java side), or do we have to pass to the reader the actual > schema/data representation? > > Thanks!
