Hi David,
Fix works fine! Thanks for that.
1 Like
Okay, running this issue with my activity journal now, too;-)
Who I import the Strava activities through the file, will activities that already are form a different source (Wahoon, HelathFit, etc.) and have a Strava activity link
- keep the data and all details from the Wahoo/etc. activity as primary data source?
- keep the Stava link into the activity there?
- be internally linked to the down/uploaded Strava activity in i.icu?
The latter is not really important, the first two points are;-)
But I do want my journal to show all (for me, mostly older) activities that came only from Strava as, e.g., Wahoo, connections were not available back then…
David, thank you for this fix! And happy new year and all the success you wish for – with i.icu, your new work setup, and in general:-)
I’ve done the Strava import, but I still have a significant number of activities that do not come through th API… because the imported once did not match, were created as duplicates, and now the original ones are till Strava links.
On example - the following are now two separate activities in i.icu from the same Strava activity:
Yes, they do have different durations:-o But pretty much the same start time…
Now, how can I consolidate that – back to “correct”?
- the original activities have my comments, RPEs, feels, etc. – but do not come through in API calls
- the newly added activities come along the API route, but do not have the info.
- I have duplicates in my history, which also change load/etc.
Those 2 have significantly different elapsed time (12m vs 15m) which is why they weren’t matched. I don’t know why the data from the Strava API is so different to the file they provided. Unfortunately you need to manually find these and copy over the feel and RPE info. Adding “Source”, “Date Created”, “Feel” and “RPE” columns to the list view is helpful.
How many of these do you have?
1 Like
But the API access to intervals is very powerful. I suppose it would be quite easy to make a python (or any other language) script that removes duplicates, based on duration and start time etc.
For example, I have one script that detects park runs, based on if the activity started on a Saturday and between 9.25 and 9.40, and it distance is between 4.9 and 5.1 km.
1 Like
Approximately 300 :-o
There are 301 pairs of activities with starting date times within 5 min, for somewhat different deltas there are (obviously) less/more. However, some of these might not even come form the Strava “all data” import, but most do.
I quickly wrote a duplicate finder on my local JSON database for my activity journal, see activity-journal/src/iicu_activity_journal/duplicates.py at main · yokuha/activity-journal · GitHub
However, I am a little hesitant to delete these activities through the API, because it would have to be fully error free, requiring quite some careful pre-operation debugging. Moreover, as of now it is (programmatically) unclear to me which one from the pair should actually be deleted. Which one is the uploaded one with the correct data or the one linked to Wahoo data/etc. Probably further issues like this to handle? Which ones?
Moreover, a number of these, maybe 50 (or some more) are activities with IDs not starting with ‘i’. However, they are real activities but their data was apparently not yet downloaded from Strava… What to do about these in the API?
Other non-i-starting entries from the API are things that do not have a direct link – such as notes/etc. – how does one distinguish these from each other?
What other pitfalls might be there?
PS: Some of the more recent duplicates also come from this problem: HealthFit sync for activities and health metrics - #97 by yokuha These ones I typically catch and manually remove while I analyze my week/etc. – which is easier (but still quite annoying;-) than wing through previous years or weeks.
The concept of anyone “owning my data” and restricting its use other than myself strikes me as extremely odd! 
I pay for strava so it’s weird to me that they make it hard for me to use my own data in whatever way I want to and using whichever other app I choose. The data comes from GARMIN anyway and is not collected by them or edited in any way for the better! F@$ck capitalism. 
Or is it just that my lactate is still high making me grumpy 
2 Likes
The data you´re getting through Strava API is modified by them. That´s what they are abusing of…
I felt so good disconnecting Intervals from Strave. I also stopped their subscription a while ago and hopefully I will be able to stop using it completely. They need to be replaced by someone who is willing to play nicer and build the community as opposed to building moats.
1 Like
Anyone can comment if the “download all data” is the original FIT file or the clobbered by Strava data fit file?
If it’s the original. Then it’s still a good archive right?
You get all the original files in that download and can do with them what you want. It is ´your´ data and Strava can´t limit anything on those.
1 Like
I ended up with duplicate activities after an import, which feels like a bug. I think I understand what happened, and may be able to fix it programatically on my end, but it seems like it might be a general issue. So worth looking into probably.
My history goes like this:
- Imported data from Strava
- Added Wahoo as source through Dropbox
- Did an export-upload of Strava
After the Wahoo data, I got some matched rides where I had dual-recorded Zwift rides on my Wahoo headunit. These weren’t 100% matches obviously, but got matched as they should.
When I did the full upload the “new” ride did not get identified as a duplicate of the original Strava activity, I’m guessing because the canonical activity data (from Wahoo) was now slightly different.
Example: i20126021 and i58098934 are both matched to the same Strava activity which seems like it shouldn’t happen after a Strava export/import.
Hope this makes sense.
Hi. I just tried to bring in all my previous strava activities - and became a supporter!
I got an email saying the import was complete, but it only added 12 activities, while it “found” over 2000…
not sure what to do now… !
Strava import complete
Intervals.icu has finished importing your Strava history.
12 activities created
Strava import for athlete i304136 from Strava | Running, Cycling & Hiking App - Train, Track & Share…
Activity column headers are en
Found 2037 activities
Date format appears to be: MMM d, yyyy, h:mm:ss a Locale: en
- Created manual i63227897 2015-07-28T05:40:00 Run Night Run, weight 72.5748 kg
Pick a couple of activities. Open the activity. Check the top right hand corner. It should say ~UPLOAD instead of ~STRAVA
This indicates that the activity has been replaced w the org file and not strava copy.
What I think happen.
2000 activities in Strava.
1988 activities in intervals.
These 1988 has been changed from STRAVA to UPLOAD
12 NEW activities were created. Somehow these are either new new or some other stuff. That you need to figure out.
Thanks!
At least one of your manual activities has elapsed time of zero. This was breaking the import. I have fixed this (these are skipped). Please give it another go. The ones already imported won’t be duplicated.
hmm… looks like I crashed and burned w/ my thinking. 
Thanks for trying to help! Much appreciated.
1 Like
I tried this, but unfortunately it does not import all of my activities. 2.158 out of 3.192.
