Create user documentation for the UniqueColumns example. (#114)
The uniqueColumns example has had little to no documentation in the past. There was no indication as to how the example should be run or even what it actually did.
This PR adds a more thorough description and gives step-by-step instructions on how to utilize the example.
diff --git a/docs/uniquecols.md b/docs/uniquecols.md
index 46b6a30..99ad701 100644
--- a/docs/uniquecols.md
+++ b/docs/uniquecols.md
@@ -17,7 +17,74 @@
# Apache Accumulo Unique Columns example
The UniqueColumns examples ([UniqueColumns.java]) computes the unique set
-of columns in a table and shows how a map reduce job can directly read a
-tables files from HDFS.
+of column family and column qualifiers in a table. It also demonstrates
+how a mapReduce job can directly read a tables files from HDFS.
+
+Create a table and add rows that all have identical column family and column
+qualifiers.
+
+```
+$ /path/to/accumulo shell -u username -p secret
+username@instance> createnamespace examples
+username@instance> createtable examples.unique
+username@instance> examples.unique> insert row1 fam1 qual1 v1
+username@instance> examples.unique> insert row2 fam1 qual1 v2
+username@instance> examples.unique> insert row3 fam1 qual1 v3
+```
+
+Exit the Accumulo shell and run the uniqueColumns mapReduce job against
+this table. Note that if the output file already exists in HDFS, it will
+need to be deleted.
+
+```
+$ ./bin/runmr mapreduce.UniqueColumns --table examples.unique --reducers 1 --output /tmp/unique
+```
+
+When the mapReduce job completes, examine the output.
+
+```
+$ hdfs dfs -cat /tmp/unique/part-r-00000
+cf:fam1
+cq:qual1
+```
+
+The output displays the unique column family and column qualifier values. In
+this case since all rows use the same values, there are only two values output.
+
+Note that since the example used only one reducer all output will be contained
+within the single `part-r-00000` file. If more than one reducer is used the output
+will be spread among various `part-r-xxxxx` files.
+
+Go back to the shell and add some additional entries.
+
+```text
+$ /path/to/accumulo shell -u username -p secret
+username@instance> table unique
+username@instance example.unique> insert row1 fam2 qual2 v2
+username@instance example.unique> insert row1 fam3 qual2 v2
+username@instance example.unique> insert row1 fam2 qual2 v2
+username@instance example.unique> insert row2 fam2 qual2 v2
+username@instance example.unique> insert row3 fam2 qual2 v2
+username@instance example.unique> insert row3 fam3 qual3 v2
+username@instance example.unique> insert row3 fam3 qual4 v2
+```
+
+Re-running the command will now find any additional unique column values.
+
+```text
+$ hdfs dfs -rm -r -f /tmp/unique
+$ ./bin/runmr mapreduce.UniqueColumns --table examples.unique --reducers 1 --output /tmp/unique
+$ hdfs dfs -cat /tmp/unique/part-r-00000
+cf:fam1
+cf:fam2
+cf:fam3
+cq:qual1
+cq:qual2
+cq:qual3
+cq:qual4
+```
+
+The output now includes the additional column values that were added during the last batch of inserts.
+
[UniqueColumns.java]: ../src/main/java/org/apache/accumulo/examples/mapreduce/UniqueColumns.java
