Skip to content

Instantly share code, notes, and snippets.

@gumdropsteve
Last active December 20, 2019 04:50
Show Gist options
  • Save gumdropsteve/386e25162061980bf1db137e97831ce8 to your computer and use it in GitHub Desktop.
Save gumdropsteve/386e25162061980bf1db137e97831ce8 to your computer and use it in GitHub Desktop.
Example of how to query your BlazingSQL logs
# This query determines the data load time and total time for all queries, showing the latest ones first.
# load time and total time being the maximum load time and total time for any node.
log_query = """
SELECT
MAX(end_time) as end_time, query_id,
MAX(load_time) AS load_time, MAX(total_time) AS total_time
FROM (
SELECT
query_id, node_id,
SUM(CASE WHEN info = 'evaluate_split_query load_data' THEN duration ELSE 0 END) AS load_time,
SUM(CASE WHEN info = 'Query Execution Done' THEN duration ELSE 0 END) AS total_time,
MAX(log_time) AS end_time
FROM
bsql_logs
WHERE
info = 'evaluate_split_query load_data'
OR info = 'Query Execution Done'
GROUP BY
node_id, query_id
)
GROUP BY
query_id
ORDER BY
end_time DESC
"""
# run query (type(log_result)==cudf.core.dataframe.DataFrame)
log_result = bc.log(log_query)
# display result
print(log_result)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment