Myria-Web

In Myria-Web, you will see an editor where you can write MyriaL or SQL queries and execute them with Myria. At the top of the screen, you should see three tabs: “Editor”, “Queries”, and “Datasets”.

Here is a quick tour of the interface:

Try the following query, which ingests a dataset from S3, stores it in the relation public:adhoc:Twitter:

Twitter = LOAD("https://s3-us-west-2.amazonaws.com/myria/public-adhoc-TwitterK.csv", csv(schema(column0:int, column1:int), skip=1));
STORE(Twitter, public:adhoc:Twitter, [$0]);

The query below computes an aggregate that it stores in a relation called public:adhoc:Followers:

Twitter = SCAN(public:adhoc:Twitter);
Followers = SELECT $0, COUNT($1) FROM Twitter;
STORE(Followers, public:adhoc:Followers, [$0]);

(The third argument in the STORE statement means to hash-partition the data on the first attribute and store it hash-partitioned across all worker instances.)

First click on “Parse”. This will show you the query plan that Myria will execute. Then click on “Execute Query”. This will execute the query and produce a link to the output.

Now select the “Profile Query” option below the query window and re-execute the query with the option ON. Below the result, you will see the “Profiling results”. Click on it. It wil show you profiling information about the way the query executed. Explore the output of the profiler.

Perfopticon

The query execution debugger is described in our paper. Please see https://idl.cs.washington.edu/papers/perfopticon/.

MyriaL

The above examples use MyriaL. For more information, please see http://myria.cs.washington.edu/docs/myrial.html.