There are no nodes on this “street”.
Thanks for reporting this!
I had a bug in my city update code that I just fixed today. I should be able to fix this up tomorrow.
Hi James, A few more streets turned up in my area recently with zero nodes:
A separate issue, but this ‘street’ is in fact in private land owned by the railway company and not accessible to the public:
Streets with no nodes (on the community these streets have no nodes, but on city level some of the streets with the same names are correct)
Here is a street I found that looks like it still has no nodes after the update. It is listed in the Riverside, IL streets list but it is a street completely in another town’s boundary so shouldn’t be here.
This street with no nodes actually has a portion within the boundary of Chicago but no nodes were created for it. It would be correct to be in the list, but with nodes.
I’m thinking that any city with a “last updated” date prior to October 15 can display this zero-node bug.
Some good news is that it looks like the system has finished its first round of updates, so Chicago should get another update pretty soon.
I have have found this street in OSM https://www.openstreetmap.org/way/7473752, that exists in CS (https://citystrides.com/streets/15497852) but with no nodes. Could that be because the street is defined as highway = living_street?
No. This is the zero-node bug from the early update process.
living_street tag is not used in my queries.
So does that mean updates proceed alphabetically?
No, it’s sorted by the last updated date - first in first out. I think the initial sort was creation date.
Another hunch: I renamed the streets awhile back (case). The nodes and street existed before in CS, but the street was now named slightly different. I think CS sees it as a new street but has not linked the nodes. Compare:
Hope that helps.
This was helpful - thanks. I manually deleted the zero-node street.
I also manually deleted all zero-node streets, to see if any get created after a ~day of updates. I now see 2 new zero-node streets, so I definitely have a lingering bug.
Ah, that explains it… No “OSM update”, but went from ~12 on the incomplete list to 2. Many were 0-node. The two remaining streets, in Keller, for me, are Jackson Road (turns private with gate) and Lambert Lane West (pavement ends and node is/was in an un-runable field) - Both have been run (to their legal/practical limits), and both have recent (~4 weeks) OSM updates from me.
Looking good James!
Hi Adamsdown https://citystrides.com/cities/130576 updated yesterday but
Watford Road https://citystrides.com/streets/15625293 a 0 node street remained.
I don’t know if it makes a difference but Adamsdown is a nested city in Cardiff, UK and Watford Road still remains in Cardiff. It’s just the node that went into Adamsdown was removed in OSM.
Hi James, for London Borough of Croydon https://citystrides.com/users/8736/cities/130937, there was an original update on 2020-09-23 which created a bunch of zero-node streets and the streets total was 2,507.
On either 10/17 or 10/18 something happened (maybe you did a manual delete?) and all those streets disappared leaving a total of 2,468 streets with only 3 to run to get 100 % so, obviously, I went straight out and ran those streets and was able to hit 100% completion (yay!)
However on 2020-10-19 Croydon got its second update and all the zero-node streets came back with a vengeance and we’re back to 2,507 total streets…
I realise you’re aware there’s still an issue in the code creating zero-node streets but thought this might add some colour. Also, I’m puzzled as to what caused the zero-node streets to disappear this past weekend as there was no update and other cities I’m running didn’t seem to have the same thing happen.
My zero-node streets are back… In Keller.
The streets are:
Liberty Hill Drive
Oakbrook Hills Ct
Magner seems to be new. 1 of 3 nodes for me. Looks like I’m heading north again. Didn’t think it was Keller, and I was right there!
These are the last streets I need to finish Keller, which looks like it got an update today, 10/20/20, so maybe it’s still updating?
Keep up the good work @JamesChevalier!