terasort.md

Apache Accumulo Terasort Example

This example uses map/reduce to generate random input data that will be sorted by storing it into accumulo. It uses data very similar to the hadoop terasort benchmark.

First, make sure the 'examples' namespace exists. If it already exists, the error message can be ignored.

$ accumulo shell -u root -p secret -e 'createnamespace examples'   

This example is run with arguments describing the amount of data:

$ ./bin/runmr mapreduce.TeraSortIngest --count 10 --minKeySize 10 --maxKeySize 10 \
    --minValueSize 78 --maxValueSize 78 --table examples.sort --splits 10

After the map reduce job completes, scan the data:

$ accumulo shell
username@instance> scan -t examples.sort
+l-$$OE/ZH c:         4 []    GGGGGGGGGGWWWWWWWWWWMMMMMMMMMMCCCCCCCCCCSSSSSSSSSSIIIIIIIIIIYYYYYYYYYYOOOOOOOO
,C)wDw//u= c:        10 []    CCCCCCCCCCSSSSSSSSSSIIIIIIIIIIYYYYYYYYYYOOOOOOOOOOEEEEEEEEEEUUUUUUUUUUKKKKKKKK
75@~?'WdUF c:         1 []    IIIIIIIIIIYYYYYYYYYYOOOOOOOOOOEEEEEEEEEEUUUUUUUUUUKKKKKKKKKKAAAAAAAAAAQQQQQQQQ
;L+!2rT~hd c:         8 []    MMMMMMMMMMCCCCCCCCCCSSSSSSSSSSIIIIIIIIIIYYYYYYYYYYOOOOOOOOOOEEEEEEEEEEUUUUUUUU
LsS8)|.ZLD c:         5 []    OOOOOOOOOOEEEEEEEEEEUUUUUUUUUUKKKKKKKKKKAAAAAAAAAAQQQQQQQQQQGGGGGGGGGGWWWWWWWW
M^*dDE;6^< c:         9 []    UUUUUUUUUUKKKKKKKKKKAAAAAAAAAAQQQQQQQQQQGGGGGGGGGGWWWWWWWWWWMMMMMMMMMMCCCCCCCC
^Eu)<n#kdP c:         3 []    YYYYYYYYYYOOOOOOOOOOEEEEEEEEEEUUUUUUUUUUKKKKKKKKKKAAAAAAAAAAQQQQQQQQQQGGGGGGGG
le5awB.$sm c:         6 []    WWWWWWWWWWMMMMMMMMMMCCCCCCCCCCSSSSSSSSSSIIIIIIIIIIYYYYYYYYYYOOOOOOOOOOEEEEEEEE
q__[fwhKFg c:         7 []    EEEEEEEEEEUUUUUUUUUUKKKKKKKKKKAAAAAAAAAAQQQQQQQQQQGGGGGGGGGGWWWWWWWWWWMMMMMMMM
w[o||:N&H, c:         2 []    QQQQQQQQQQGGGGGGGGGGWWWWWWWWWWMMMMMMMMMMCCCCCCCCCCSSSSSSSSSSIIIIIIIIIIYYYYYYYY

Of course, a real benchmark would ingest millions of entries.