analytic Coffee Lovers,
We will meet in Lean Coffee Style Online on Friday 5/3/2019 @ 12:00noon EDT to discuss all things Tessitura Analytics, analytics & data science within the Tessitura Community.
The Agenda of our conversation will be set by the folks who show up related to our topic of interest. I've already set our lean coffee Table web site for this meeting.
We will pick who goes first and how long we want to continue on a particular topic with Lean Coffee rules.
We are going to use Google Meet for our session. The link will be https://meet.google.com/riv-wjpp-rjk.
Please try to Join from a computer with a camera so that we can all see one another. If you must join from a phone:
Phone Numbers (US) +1 629-888-0820 PIN: 137 807 574#
The meet.google.com link will include a full list of phone numbers for all supported countries.
You will have a much better experience on this call if you can fully participate. Having a webcam, and computer audio link seems to make things better. Please feel free to reach out to me with any challenges you might be having in connecting into our conversations. Or join in on this discussion going on right now in our group at https://community.tessituranetwork.com/topical_groups/analytics-coffee/f/discussions/22402/connecting-to-analytic-coffee
Here are the Notes from today's conversation. Thanks all for joining.
analyticCoffee-03May2019.pdf
Just wanted to share my joy. I've been plugging away at the hockeystick graph and I'm starting to crack it using the new BI.VTs and this little piece of code I cooked up. I'm overly proud of it due to being a self taught SQL novice. Looks promising so far. Thanks for the tips all
declare @lastperf datetime select @lastperf = MAX(perf_dt) from BI.VT_ORDER_DETAIL where prod_season_no = 553
Select days_prior, sum(sales) over (order by days_prior rows unbounded preceding)
from ( select (DATEDIFF(DD, @lastperf, order_dt)) as days_prior, sum(paid_amt) sales from BI.VT_ORDER_DETAIL where prod_season_no = 553
group by (DATEDIFF(DD, @lastperf, order_dt))
) t
order by days_prior asc
thanks for sharing. I'm finding opening the table BI.VT_ORDER_DETAIL very very very very slow. What type of performance are you seeing from this view? Just to do a row count,
select count(*) from BI.VT_ORDER_DETAIL
in my RAMP test environment, it is taking an impossibly long amount of time.....
Like I waited more than 20 minutes before sending this note with no results.
Chris Wallingford any idea why this is so slow to use? I'm sure that this table has a lot of rows in it. We do at least 1 GA hours of 7000 seats every day for 5-6 years. If I understand what is going on here that means that from just our GA House records I have more than 13 million rows in this very complicated view. Are there any performance optimizations available for this view? I think I read about ways to not show blacked out seats or unsold seats...
After 32 minutes, I got a row count of 10,368,517. So clearly some data trimming is already happening.
Hi Tom,
It is expected that a FULL load of the source as you've queried it, into your Analytics warehouse might the longest running of all the sources. The extraction is pulling all SLIs (not all seats) and is highly optimized. BI.VT_DW_PERFORMANCE_SEATING pulls inventory, so those GA houses you mentioned. Blackout seats are excluded by default via a T_DEFAULTS setting. Applying the Suppress Unsold inventory keyword to GA houses where a measure of capacity isn't important is recommended.
Best,Chris
Does the Keyword trim the view?
Yes. Here are the details:
https://www.tessituranetwork.com/Help_System_v150/Content/Tessitura%20Analytics%20Tech/Suppress%20Unsold%20TKW.htm
Cheers,Chris
What happens if
"Prior to running the Load Data Warehouse job for the first time, add the Suppress Unsold keyword to all appropriate production elements." One did not add the keywords?
Is there a process to apply these changes after the Warehouse has been put into operations?
P.S. I can open a TASK ticket when we are ready to look at this in detail.