Data Panel not sorting by name

[daveN69RZ]

As per the screenshot below, I'm not getting a correctly sorted list of drawings by name.  It's more than a bit inconvenient as I've got close to 100 parts/assemblies to draw up and need to be able to check that I've captured all of them easily. It's happening in other folders.

[matt.pooley]

Hi Dave,

After digging into this for a bit I believe I understand what is happening, and while it does make sense, it doesn't help your issue. The problem is that our logic for sorting is trying to be smart when it sees numbers, for example it is trying to make sure that 99 comes before 100 when sorted by name.

Mapping that logic to the data in your picture, it is ignoring the space characters in the first numeric section of the names and seeing them as :

• 05111(DS1...)
• 05172(TB1...)
• 05182(TB2...)
• 05191(DS2...)
• 05622(DS1...)
• 052019(TB3...)
• 052219(CB2...)

What this means is that when some files have a longer sequence of numbers at the start of their names, they will head towards the bottom of the numeric section of the list (since they are seen as larger values).

Hopefully this explains what is happening, whether what is happening is 'right' is a different question, I have seen that we do stop parsing numeric values for a range of other seperator characters (e.g. '-'), possibly we should also stop once we see a space character such as you have in your names.

Regards,

Matt Pooley

On Behalf of the Fusion Dev Team

[daveN69RZ]

Could you share a list of separaters? I've 'fixed' (for a given value of fixed) the issue by inserting a '-' between reference and description.

If it's a little fix at your end, then by all means add a space to the list. If it's just me, then feel free to ignore it. After all what is sensible to one engineer is never sensible to another. 

[matt.pooley]

Hi Dave,

Doing some more digging (into the code this time), there is some relatively complex processing that strips out space characters and then splits the remaining text into segments of alphanumeric characters. The segments are then compared one at a time from left to right to see which order to sort things. this means that anything other than space, a-z, A-Z and 0-9 is treated as a separator.

A change to not strip spaces (and instead use them as another divider between segments) is possible, though I'm nervous about how many people might be disrupted by that change, I'll aim to add an option or text command that will let people override the behavior for spaces without everyone seeing a difference...

Regards,

Matt.

[Witsend3486]

Matt,

I've read your sorting algorithm description and a) I don't entirely understand it and b) per the following screenshot, I don't see that it's working...

unless it is stopping its sorting effort once it sees a transition from numeric to alphabetic characters, or whatever.

Since there is already another sorting option listed in the drop down, "Sort by Last Updated", could you maybe just add a third option of "Stupid Sort by Name" that just sorts by the raw character string?  I suppose this will require a major decree from on high and, possibly, a full corporate restructuring with a PR campaign, but maybe?

Respectfully,

Bill

[matt.pooley]

Hi Bill,

I took another look into this, and it turns out there was an unexpected behavior change in the May release. What is now happening is the names are being split into chunks of alpha-numeric characters and then sorted either numerically or alphabetically depending on whether each chunk starts with a number or a letter. Applying that logic to your data, it is seeing your (chunked) names as :

270CA28945_20190823_00_00 -> 270CA28945 : 20190823 : 00 :00

270CA30220_20190823_00_00 -> 270CA30220 : 20190823 : 00 : 00

270CA28944_20190823_00_00 -> 270CA28944 : 20190823 : 00 : 00

The sort is then treating all these chunks as numbers (by truncating at any non-numeric characters), which is turning them all into exactly the same data for sorting - 270 : 20190823 : 00 : 00

There is still something missing in my understanding though, since it should be possible to modify the last chunk of these names and see sorting happening, that wasn't the case when I tested, so I'm continuing to explore.

We are working to fix the implementation and make it more reliable, though it won't be in time for the next release (which is soon).

Thank you to Bill for the nudge, I hadn't realized this change had been made...

Regards,

Matt

[kcopage]

Hi Matt , 

I am seeing the same issue, and it is extremely frustrating. This thread was from 2019, was it ever completed? The web view and desktop connector sort things just fine. 

[matt.pooley]

Hi,

I haven't dug into this area for a few years so not sure if the change from 2019 was incomplete or if there has been a more recent regression in behavior. Will pass this to the relevant group and see what might be happening.

Regards,

Matt.

[kcopage]

Thank you, It is at least helpful to understand how/why it is sorting things the way it is... In my case our parts have an item number followed by a description, sometimes the description starts with a number which is where the issue was.
Someting like the example bellow would sort incorrectly:
30004P505 Part A
30004P504 3D print Part B

Now FINALLY understanding that it ignores spaces while sorting... Never seen a sorting metric behave this way.
Anywho, replacing the space with an underscore is a stop-gap solution.

30004P504_3D print Part B
30004P505 _Part A

I tried writing an API script to adjust all of our item numbers in this manner (we have thousands) however it seems the API is not able to re-name design files.

Eager for your feedback, thanks!