Github user denalex commented on a diff in the pull request:
https://github.com/apache/incubator-hawq/pull/1344#discussion_r171036634
--- Diff: pxf/pxf-ignite/README.md ---
@@ -0,0 +1,67 @@
+# Accessing Ignite database using PXF
+
+The PXF Ignite plug-in enables to access the [Ignite
database](https://ignite.apache.org/) (both read and write operations are
supported) via REST API.
+
+
+# Prerequisites
+
+Check the following before using the plug-in:
+
+* The Ignite plug-in is installed on all PXF nodes;
+
+* The Ignite client is installed and running at the `IGNITE_HOST`
(`localhost` by default; see below), and it accepts http queries from the PXF
(note that *enabling Ignite REST API does not require changes in Ignite
configuration*; see the instruction on how to do that at
https://apacheignite.readme.io/docs/rest-api#section-getting-started).
+
+
+# Syntax
+
+```
+CREATE [READABLE] EXTERNAL TABLE <table_name> (
+ <column_name> <data_type>[, <column_name> <data_type>, ...] | LIKE
<other_table>
+)
+LOCATION
('pxf://<ignite_table_name>?PROFILE=Ignite[&<extra-parameter>&<extra-parameter>&...]')
+FORMAT 'CUSTOM' (formatter='pxfwritable_import');
+```
+where each `<extra-parameter>` is one of the following:
+* `IGNITE_HOST=<ignite_host_address_with_port>`. The location of Ignite
client node. If not given, `127.0.0.1:8080` is used by default;
+* `IGNITE_CACHE=<ignite_cache_name>`. The name of Ignite cache to use. If
not given, this parameter is not included in queries from PXF to Ignite, thus
Ignite default values will be used (at the moment, this is `Default` cache).
This option is **case-sensitive**;
+* `BUFFER_SIZE=<unsigned_int>`. The number of tuples send to (from) Ignite
per a response. The same number of tuples is stored in in-plug-in cache. The
values `0` and `1` are equal (cache is not used, each tuple is passed in it's
own query to Ignite). If not given, `128` is used by default;
+* `PARTITION_BY=<column>:<column_type>`. See below;
+* `RANGE=<start_value>:<end_value>`. See below;
+* `INTERVAL=<value>[:<unit>]`. See below.
+
+
+# Partitioning
+## Introduction
+
+PXF Ignite plugin supports simultaneous access to Ignite database from
multiple PXF segments. *Partitioning* should be used in order to perform such
operation.
+
+If the partitioning is not used, all the data will be retrieved by a
single PXF segment.
+
+Partitioning in PXF Ignite plug-in works just like in PXF JDBC plug-in.
--- End diff --
since this is a separate plugin and JDBC partitioning may change
independently, it would be good to describe how it woks for Ignite here.
---
