Reputation: 5157
A forewarning: I am not a SQL Dev. I am looking at this query and trying to modify it to resolve an issue with the results.
I have the follow query which works (syntactically):
SELECT
MIN(issue_time) AS incident_time,
MIN(issue_status) AS issue_status,
MIN(monitor_start_time) AS monitor_start_time,
MAX(resolved_time) AS resolved_time,
incident_id,
MAX(model_handle) AS model_handle,
device_retrieval_id,
MAX(last_status_time) AS last_status_time,
MAX(last_analysis_time) AS last_analysis_time,
n.node_name,
MAX(last_problem_time) AS last_problem_time,
SUM(problem_count) AS problem_count
FROM (
SELECT
issue_time, issue_status, monitor_start_time, resolved_time, incident_id,
model_handle, device_retrieval_id, last_status_time, last_analysis_time,
last_problem_time, problem_count
FROM x_network_device_node_status
UNION
SELECT
issue_time, issue_status, monitor_start_time, resolved_time, incident_id,
model_handle, device_retrieval_id, last_status_time, last_analysis_time,
last_problem_time, problem_count
FROM x_network_device_node_status_hist
WHERE incident_id IN (
SELECT DISTINCT incident_id
FROM (
SELECT incident_id
FROM x_network_device_node_status
UNION
SELECT incident_id
FROM x_network_device_node_status_hist
WHERE issue_time >= '2016-12-10'
) AS i
)
) AS x LEFT JOIN
x_nodes AS n ON x.device_retrieval_id = n.node_retrieval_id
GROUP BY incident_id, device_retrieval_id, node_name
ORDER BY incident_time DESC;
I recently realized that that we should not be selecting the MAX(resolved_time) AS resolved_time
, but instead should be getting the resolved_time
for each incident_id
from the record with the most recent issue_time
.
For example, if the UNION of records in the database for a particular incident_id was:
╔═════════════════════════╦══════════════╦═════════════════════════╦═════════════════════════╦═════════════════════╦══════════════╦═════════════════════╦═════════════════════════╦═════════════════════════╦═════════════════════════╦═══════════════╗
║ issue_time ║ issue_status ║ monitor_start_time ║ resolved_time ║ incident_id ║ model_handle ║ device_retrieval_id ║ last_status_time ║ last_analysis_time ║ last_problem_time ║ problem_count ║
╠═════════════════════════╬══════════════╬═════════════════════════╬═════════════════════════╬═════════════════════╬══════════════╬═════════════════════╬═════════════════════════╬═════════════════════════╬═════════════════════════╬═══════════════╣
║ 2016-12-13 17:15:00.000 ║ 1 ║ 2016-12-13 17:46:28.000 ║ 2016-12-13 18:16:34.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-13 18:16:34.000 ║ 2016-12-13 18:01:28.000 ║ NULL ║ 0 ║
║ 2016-12-13 18:15:00.000 ║ 1 ║ NULL ║ 2016-12-13 18:31:28.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-13 18:31:28.000 ║ NULL ║ NULL ║ 0 ║
║ 2016-12-13 18:31:28.000 ║ 2 ║ 2016-12-14 09:16:28.000 ║ 2016-12-15 08:31:30.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 08:31:30.000 ║ 2016-12-15 08:16:28.000 ║ 2016-12-15 08:16:28.000 ║ 6 ║
║ 2016-12-15 08:30:00.000 ║ 1 ║ 2016-12-15 10:01:31.000 ║ 2016-12-15 11:46:28.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 11:46:28.000 ║ 2016-12-15 11:31:31.000 ║ 2016-12-15 11:31:31.000 ║ 3 ║
║ 2016-12-15 11:45:00.000 ║ 1 ║ NULL ║ 2016-12-15 12:01:31.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 12:01:31.000 ║ NULL ║ NULL ║ 0 ║
║ 2016-12-15 12:00:00.000 ║ 1 ║ 2016-12-15 12:16:28.000 ║ 2016-12-15 12:31:29.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 12:31:29.000 ║ NULL ║ NULL ║ 0 ║
║ 2016-12-15 12:30:00.000 ║ 1 ║ NULL ║ 2016-12-15 13:01:31.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 13:01:31.000 ║ NULL ║ NULL ║ 0 ║
║ 2016-12-15 13:01:31.000 ║ 2 ║ 2016-12-15 16:46:28.000 ║ NULL ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 22:31:28.000 ║ 2016-12-15 22:31:28.000 ║ 2016-12-15 17:31:28.000 ║ 2 ║
╚═════════════════════════╩══════════════╩═════════════════════════╩═════════════════════════╩═════════════════════╩══════════════╩═════════════════════╩═════════════════════════╩═════════════════════════╩═════════════════════════╩═══════════════╝
I would like the result to be:
╔═════════════════════════╦══════════════╦═════════════════════════╦═══════════════╦═════════════════════╦══════════════╦═════════════════════╦═════════════════════════╦═════════════════════════╦════════════╦═════════════════════════╦═══════════════╗
║ incident_time ║ issue_status ║ monitor_start_time ║ resolved_time ║ incident_id ║ model_handle ║ device_retrieval_id ║ last_status_time ║ last_analysis_time ║ node_name ║ last_problem_time ║ problem_count ║
╠═════════════════════════╬══════════════╬═════════════════════════╬═══════════════╬═════════════════════╬══════════════╬═════════════════════╬═════════════════════════╬═════════════════════════╬════════════╬═════════════════════════╬═══════════════╣
║ 2016-12-13 17:15:00.000 ║ 1 ║ 2016-12-13 17:46:28.000 ║ NULL ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 22:31:28.000 ║ 2016-12-15 22:31:28.000 ║ MyNodeName ║ 2016-12-15 17:31:28.000 ║ 11 ║
╚═════════════════════════╩══════════════╩═════════════════════════╩═══════════════╩═════════════════════╩══════════════╩═════════════════════╩═════════════════════════╩═════════════════════════╩════════════╩═════════════════════════╩═══════════════╝
But instead, it is coming back as:
╔═════════════════════════╦══════════════╦═════════════════════════╦═════════════════════════╦═════════════════════╦══════════════╦═════════════════════╦═════════════════════════╦═════════════════════════╦════════════╦═════════════════════════╦═══════════════╗
║ incident_time ║ issue_status ║ monitor_start_time ║ resolved_time ║ incident_id ║ model_handle ║ device_retrieval_id ║ last_status_time ║ last_analysis_time ║ node_name ║ last_problem_time ║ problem_count ║
╠═════════════════════════╬══════════════╬═════════════════════════╬═════════════════════════╬═════════════════════╬══════════════╬═════════════════════╬═════════════════════════╬═════════════════════════╬════════════╬═════════════════════════╬═══════════════╣
║ 2016-12-13 17:15:00.000 ║ 1 ║ 2016-12-13 17:46:28.000 ║ 2016-12-15 13:01:31.000 ║ 1427068756009427097 ║ 85983911 ║ 332265337 ║ 2016-12-15 22:31:28.000 ║ 2016-12-15 22:31:28.000 ║ MyNodeName ║ 2016-12-15 17:31:28.000 ║ 11 ║
╚═════════════════════════╩══════════════╩═════════════════════════╩═════════════════════════╩═════════════════════╩══════════════╩═════════════════════╩═════════════════════════╩═════════════════════════╩════════════╩═════════════════════════╩═══════════════╝
Upvotes: 0
Views: 57
Reputation: 131
SELECT
MIN(issue_time) AS incident_time,
MIN(issue_status) AS issue_status,
MIN(monitor_start_time) AS monitor_start_time,
(SELECT TOP 1 xxx.resolved_time
FROM
(SELECT xx.issue_time,xx.resolved_time
FROM x_network_device_node_status xx
WHERE xx.incident_id=x.incident_id
UNION ALL
SELECT xx2.issue_time,xx2.resolved_time
FROM x_network_device_node_status_hist xx2
WHERE xx2.incident_id=x.incident_id) AS xxx
ORDER BY xxx.issue_time DESC) AS resolved_time,
incident_id,
Upvotes: 1
Reputation: 1
I'm not sure about the syntax, but the idea is to use CTE and query it.
;WITH CTE AS
(SELECT
issue_time, issue_status, monitor_start_time, resolved_time, incident_id,
model_handle, device_retrieval_id, last_status_time, last_analysis_time,
last_problem_time, problem_count
FROM x_network_device_node_status
UNION
SELECT
issue_time, issue_status, monitor_start_time, resolved_time, incident_id,
model_handle, device_retrieval_id, last_status_time, last_analysis_time,
last_problem_time, problem_count
FROM x_network_device_node_status_hist
WHERE incident_id IN (
SELECT DISTINCT incident_id
FROM (
SELECT incident_id
FROM x_network_device_node_status
UNION
SELECT incident_id
FROM x_network_device_node_status_hist
WHERE issue_time >= '2016-12-10'
) AS i
)
)
SELECT
MIN(issue_time) AS incident_time,
MIN(issue_status) AS issue_status,
MIN(monitor_start_time) AS monitor_start_time,
(SELECT resolved_time FROM CTE WHERE issue_time = (SELECT MAX(issue_time) FROM CTE)) AS resolved_time,
incident_id,
MAX(model_handle) AS model_handle,
device_retrieval_id,
MAX(last_status_time) AS last_status_time,
MAX(last_analysis_time) AS last_analysis_time,
n.node_name,
MAX(last_problem_time) AS last_problem_time,
SUM(problem_count) AS problem_count
FROM CTE AS x LEFT JOIN
x_nodes AS n ON x.device_retrieval_id = n.node_retrieval_id
GROUP BY incident_id, device_retrieval_id, node_name
ORDER BY incident_time DESC;
Upvotes: 0