[
https://issues.apache.org/jira/browse/SOLR-15197?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joel Bernstein updated SOLR-15197:
----------------------------------
Description:
The initial implementation of temporal graph queries supported ten second time
windows which is useful for log analytics use cases. This ticket will allow for
event correlation across daily windows, which is useful for many different
fields.
This ticket will add the support for daily windows and add the syntax to
support other time windows as well. Follow-on tickets will add more time
windows.
Below is the sample syntax for temporal graph query with a window of 1 DAY.
{code:java}
nodes(daily_stock_returns,
search(daily_stock_returns,
q="ticker_s:jpm AND change_d:[2 TO *]",
fl="day_s",
sort="change_d desc",
rows="50")
walk="day_s->day_s",
gather="ticker_s",
fq="close_d:[1 TO *]",
window="1DAY",
lag="2",
count(*))
{code}
What this query does is count all tickers that are up by atleast 1, in a one
day window, 2 days before the ticker jpm rises by atleast 2. This surfaces
which tickers most frequently rise two days before jpm rises.
This demonstrates event correlation is stock market data.
was:
The initial implementation of temporal graph queries supported ten second time
windows which is useful for log analytics use cases. This ticket will allow for
event correlation across daily windows, which is useful for many different
fields.
This ticket will add the support for daily windows and add the syntax to
support other time windows as well. Follow-on tickets will add more time
windows.
Below is the sample syntax for temporal graph query with a window of 1 DAY.
{code:java}
nodes(daily_stock_returns,
search(daily_stock_returns,
q="ticker_s:jpm AND change_d:[2 TO *]",
fl="day_s",
sort="change_d desc",
rows="50")
walk="day_s->day_s",
gather="ticker_s",
fq="close_d:[1 TO *]",
window="1DAY",
lag="2",
count(*))
{code}
What this query does is count all tickers that are up by atleast 1, in a one
day window, 2 days before the ticker jpm rises by atleast 2. This surfaces
which tickers most frequently rise two days before jpm rises.
> Support temporal graph queries with daily windows
> -------------------------------------------------
>
> Key: SOLR-15197
> URL: https://issues.apache.org/jira/browse/SOLR-15197
> Project: Solr
> Issue Type: Improvement
> Reporter: Joel Bernstein
> Assignee: Joel Bernstein
> Priority: Major
> Fix For: main (9.0)
>
> Attachments: SOLR-15197.patch
>
>
> The initial implementation of temporal graph queries supported ten second
> time windows which is useful for log analytics use cases. This ticket will
> allow for event correlation across daily windows, which is useful for many
> different fields.
> This ticket will add the support for daily windows and add the syntax to
> support other time windows as well. Follow-on tickets will add more time
> windows.
> Below is the sample syntax for temporal graph query with a window of 1 DAY.
> {code:java}
> nodes(daily_stock_returns,
> search(daily_stock_returns,
> q="ticker_s:jpm AND change_d:[2 TO *]",
> fl="day_s",
> sort="change_d desc",
> rows="50")
> walk="day_s->day_s",
> gather="ticker_s",
> fq="close_d:[1 TO *]",
> window="1DAY",
> lag="2",
> count(*))
> {code}
> What this query does is count all tickers that are up by atleast 1, in a one
> day window, 2 days before the ticker jpm rises by atleast 2. This surfaces
> which tickers most frequently rise two days before jpm rises.
> This demonstrates event correlation is stock market data.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
