CSV benchmarks

README.md

@@ -133,8 +133,8 @@ Hardware:
 |--------------------|-------------------|------------|-------------------|
 | Simple Aggregation | 36,000 msgs / sec | 256 MiB    | 102 MiB           |
 | Enrichment         | 13,000 msgs /sec  | 368 MiB    | 124 MiB           |
-| CSV Disk           |                   |            |                   |
-| CSV Memory         |                   |            |                   |
+| CSV Disk Join      | 11,500 msgs /sec  | 312 MiB    | 152 MiB           |
+| CSV Memory Join    | 33,200 msgs / sec | 300 MiB    | 107 MiB           |
 
 ### Simple Aggregate 
 
@@ -158,5 +158,19 @@ python3 cmd/publish-test-data.py --num-messages=1000000 --topic="topic-enrich"
 /usr/bin/time -l python3 cmd/sql-flow.py run /Users/danielmican/code/github.com/turbolytics/sql-flow/dev/config/benchmarks/enrich.yml
 ```
 
+### CSV Disk Join
+
+```
+python3 cmd/publish-test-data.py --num-messages=1000000 --topic="topic-csv-filesystem-join"
+SQLFLOW_STATIC_ROOT=/Users/danielmican/code/github.com/turbolytics/sql-flow/dev /usr/bin/time -l python3 cmd/sql-flow.py run /Users/danielmican/code/github.com/turbolytics/sql-flow/dev/config/examples/csv.filesystem.join.yml
+```
+
+## CSV Memory Join
+
+```
+SQLFLOW_STATIC_ROOT=/Users/danielmican/code/github.com/turbolytics/sql-flow/dev /usr/bin/time -l python3 cmd/sql-flow.py run /Users/danielmican/code/github.com/turbolytics/sql-flow/dev/config/examples/csv.mem.join.yml
+python3 cmd/publish-test-data.py --num-messages=1000000 --topic="topic-csv-mem-join"
+```
+
 --- 
 Like SQLFlow? Use SQLFlow? Feature Requests? Please let us know! [email protected]

cmd/publish-test-data.py

@@ -32,6 +32,14 @@
   "originalTimestamp": "2015-12-12T19:11:01.152Z"
 }
 
+cities = [
+    'San Fransisco',
+    'Baltimore',
+    'New York',
+    'Miami',
+    'Asheville',
+]
+
 
 @click.command()
 @click.option('--num-messages', default=1001, type=int)
@@ -45,7 +53,7 @@ def main(num_messages, topic):
     producer = Producer(conf)
     for i in range(num_messages):
         e = copy.deepcopy(event)
-        e['properties']['city'] = e['properties']['city'] + str(random.randrange(0, 1000))
+        e['properties']['city'] = random.choice(cities)
         j_event = json.dumps(e)
         producer.produce(topic, value=j_event)
         if i % 1000 == 0: