6 Replies Latest reply: Mar 14, 2012 10:57 PM by distri
distri Level 1 Level 1



   today I've downloaded again all raw daily logs now in format tsv.


  I've discovered that they have  with multiple errors like field shifted and wrong characters  ( i.e instead of ü I've found square root symbols).


The good point is that the new reports look to have more records per day compared to the old reports: old reports had missing record.





  • Darudh Level 1 Level 1



    can confirm this. In addition the Storefront_id is at times missing or invalid ( 4 or 5 digits only )



  • Bowie Poon Level 1 Level 1

    Same here.

    Data fields are shifted and for some records, the URL values are somehow truncated and append to the values in title field, while leaving the URL field empty.


    Is someone working on this?

  • distri Level 1 Level 1

    I can confirm there are some random errors like truncations, fields shifted or missing.

    Another problem is that if you download a daily raw log just the day after in the early morning , records reported stops a couple of hours before midnight. I'll explain better with an example:


    today is the early morning of the 9 march and I download the raw report of the 8 march: looking at the timestamp

    field the last record reported is around the 21PM; if I reload the same raw log of the 8 march in the evening of the 9 march now the last record timestamp is around midnight.

    It looks like the raw report is published and downloadable, but still in progress.


    To analyze and get statistic from the raw reports I've created a Filemaker database, so it's easier to detect errors.


    Another thing of the new reporting that it should be improved is the storefront because it does not report the real geographical location of the iTunes U user. There are users from different countries that select other or wrong iTunes Client Country Store instead of the real geographical locations. I've compared this using the IP number field of the reports and confirmed by the daily log raw reports that are still available from the iTunes U "dz" sites via "GetDailyReportLogs" web service ( see iTunes U Administration Guide for Private Site).

    I suppose that storefront is used instead of the real IP address because not all countries ara available in iTunes U Store or probably for privacy reasons, but it would be nice if somebody could confirm this.





  • distri Level 1 Level 1

    I reply to myself about the geolocation and Storefront: first two octets of the IP address are enough to get a quite good geolocation.



  • Bowie Poon Level 1 Level 1

    The random errors like truncations, fields shifted or missing, messy data in the report is really a headache to us. We are unable to generate a summarised report base on those messy data.


    Say, one of the title value is 'Basic Chinese 1". In the report, some of the title values becomes:

    Basic Chinese 1ft Windows XP Professional Service Pack 3 (Build 2600)) AppleWebKit/534.54.16
    Basic Chinesele.com/WebObjects/DZR.woa/wa/viewPodcast?cc=us&id=466689031
    Basic Chinese 1es-iPhone/5.0.1 (3; 32GB) iTunesU/1.1

    There is no way we could identify such error without eyeball checking.

    Besides, we can't identify the item being downloaded/browsed using the ITUNES_ID either, since they are different (sometimes).


    To Apple Support, please fix it as soon as possible. Or could you resume the weekly summary report email before problem solved??

  • distri Level 1 Level 1



      I'm reading you are using ITUNES_ID to identify if item being downloaded/browsed.

      I'm interested to know a good way to create summary reports using raw logs. I'm using a Filemaker database;to create cross reports I mainly use TITLE, EPISODE_TITLE, and ACTION_TYPE fields.

      I I'm interested to know something more about ITUNES_ID and EPISODE_ID fields, because I did not find any good explanation about those.


       Any suggestion is welcome.