see also / based on : /wiki/spaces/KASRLCS/pages/2777088016
Objective | special exports for aggregated data in customized format |
Start Date | 07.06.2021 |
Due date |
|
Delivery Date to customer | 07.06.2021 |
Key outcomes | aggregated data export |
Status | delivered |
JIRA | IAM-6057, IAM-6124 https://kantarware.visualstudio.com/KT-KantarTagging/_workitems/edit/977984 |
Dependencies | /wiki/spaces/RESEARCH/pages/1538916548 Azure feature: 847806 |
export data for site
tv2stream
for one day
Emedia is procduced with flag “de.spring.mapreduce.streaming.processPlayStates.version=V2”
(= The so called “1s-patch” is included and active)
Result is one export with two output stages
Filename: /user/tns-no/ucdw/public/tns-nosftp/api/${day}_tv2_stream.txt (tab separated)
Stack No1
cq is empty
stream is aggregated over every hierarchie (upper hierarchie includes the lowers too)
view is combination of client,site,uid
start the hierarchy with “tv2/”
maximum number of hierarchy is set to 6 (start with 0, f.e. sumo/live/2/3/4/5/6asdsad)
Stack No2
same as stack-no-1, but
cq is filled
Scope
The specification from the customer is as follows
Specification for FTP-delivery from TNS Scores to TV 2 V1.1.
Table structure:
Field name | Type | Comment |
Stack_NO | INTEGER | (1,2) |
Program-id | STRING |
|
Stream | STRING |
|
Content type | STRING |
|
Total_Playtime | FLOAT |
|
Total_Playtime_Client | FLOAT |
|
Views | FLOAT |
|
Views_client | FLOAT |
|
Clients | FLOAT |
|
Date | DATE |
|
All fields are as defined in https://scores.tns-cs.net/ report generation tool.
Naming convention for field names is as used in http://scores.tns-cs.net
CQ (program-id)
CT (Content Type)
Streampath (Stream)
Stack_NO:
Stack_NO is a generated field that divides two parts of the delivery.
It can either be 1 or 2:
Stack_NO: 1 | Stack_NO: 1 We want the unique clients and usage numbers for each node in the stream-parameter starting from the node UPDATE: changed to tv2/* in March 2022 https://kantarware.visualstudio.com/KT-KantarTagging/_workitems/edit/977984 Program-id-parameter: No PROGRAM-ID-parameter.
Stream-parameter: /tv2/live/tv_2_val/linad/ It should be included in the following Stream -nodes: /tv2/ /tv2/live/ /tv2/live/tv_2_val / /tv2/live/tv_2_val /linad/
Content type-parameter: It should be divided by Content type at the deepest available node.
This part of the dataset should be marked with
|
Stack_NO: 2 |
Stack_NO: 2 We want the unique clients and usage numbers for each node in the stream-parameter starting from the node UPDATE: changed to tv2/* in March 2022 https://kantarware.visualstudio.com/KT-KantarTagging/_workitems/edit/977984 Program-id-parameter: With program-id-parameter for all program-id with registered viewing.
Stream-parameter: Most detailed stream-parameter
Most detailed content type-parameter. (Including versions for content type)
|
Expected outcome
Export example / description
Based on the request “Specification for FTP-delivery from TNS Scores to TV 2 V1.1” (see below)
Delivery-frequency: | daily |
Timerange: |
Localtime Norway 03:00 - 03:00 UPDATE: changed in March 2022 https://kantarware.visualstudio.com/KT-KantarTagging/_workitems/edit/977984 |
Format | Txt, tab separated |
Files per day | 1 |
Location |
The export is following the specification above.
Values “playtime per client” and “views per client” are rounded to whole numbers: <0,5>
Stack 1: contains total values for higher-level sections and streams (the entire directory tree in Scores starting with “tv2/”, line 5 in the example below)
Stack 2: contains total values for single streams (=per cq-variable)
Stack 1 and Stack 2 are separated by all available ct-values
(formerly “sumo/live” is changed to “tv2/live” in July 2021, therefore the screenshots below are a bit outdated, another change in March 2022 to start with “tv2/”)
Success metrics
The numbers and values are comparable with the results in Scores
All available cq-values and ct-values are reflected
The export is live since 06.2021