Export and Import Ranks

[Edward Greathouse]

My company uses a project within JIRA as a template for other projects. Whenever a project manager (who is also a JIRA admin) wants to create a new project, they export all the issues of a particular Epic within the template project to a CSV file and import it using JIRA's CSV importer tool. Unfortunately, the "Rank" of the issues (therefore the order in which the issues are listed in the backog) are not maintained (they appear in the order of their key). I checked the filter for the board that the backlog is using and it says "project = TBDCOM2 ORDER BY Rank ASC". 

When I check the value of "Rank", the format is a little bizarre to me ("0|i013yk:", "0|i013yf:", etc.). I tried changing this value in the CSV file to be more meaningful like "1", "2", "3", etc. but the backlog still did have order I desired.

What else can we do to maintain the order of the backlog when exporting from our "template" project and importing to a new project? We don't want to spend hours reorganizing the order of the backlog each time a new project is created. 

[Andy Heinzer]

The rank values are using a system called lexoranking.  This lets Jira Software rank issues using an alphanumeric system.   You can't change these into purely numeric values and expect Jira to be able to use these.  But you can include the existing Rank values when you export this issues into a CSV file.  In order to do that, you just need to make sure that you include the rank value on the issue navigator if you are exporting the current fields.

What is more, you can also import these rank values as they exist into the new Jira instances as well.  You just need to map the Rank field during the CSV import.  This way the issues will maintain their rank values in relation to each other at the time the export was created.

There was a related case to this over in the thread https://community.atlassian.com/t5/Jira-Software-questions/Why-doesn-t-the-query-to-search-a-Sprint-s-contents-show-in/qaq-p/679714

Basically, you need to include the Rank field when you create the CSV exports, to be able to maintain the order that issues appear on a board/backlog.

[Clinton]

Hi Andrew,

If you exported issues and used the Rank field in Excel and re-imported back to Jira are there any known risks. Can you expand on this sentence, please?

Basically, you need to include the Rank field when you create the CSV exports, to be able to maintain the order that issues appear on a board/backlog.

Such as, could this a copy paste or format error with a "Rank" field or a change to an existing rank field 'break/ alter' the ranking of the other issues when importing the exported data via csv. back to Jira?

Appreciate your insight, please.

[Andy Heinzer]

CSV imports have the ability to overwrite existing issue data.   But only happens when you import and map the issue key field, if that issue key already exists in this Jira instance, the data you import gets updated, instead of creating a new issue.

If you are not mapping the issuekey field, or the value in the issuekey field does not exist on the destination Jira site, new issues get created for that data.  In this scenario there is way this data import will change other issue values.

I'm not sure I understand how this import would break the ranking system.  I suppose it could be possible in some scenarios, but I don't know of any off the top of my head that would break the ranking system.  

One scenario that comes to mind is what happens if you import an issue with a specific rank value that another issue is already using?   This can cause duplicate rank values to appear, but this condition doesn't cause ranking to break, it only could cause confusion between the two issues as to understand which should be ranked higher/lower in relation to each other.  If there are no boards where both such issues appear, I doubt you would ever notice this.   But if there are, then there are other steps you can take to correct this duplicate value such as that described in the KB How To Fix Duplicate Rank Values For a Rank Field

[Clinton]

Hi Andrew,

Thank you, this is exactly the clarification I required and now I have a much better understating of Jira's behaviour with csv. imports and Ranking.

Thank you!

[David Stokes]

Hi

I've got a set of issues I want to be ordered in a particular way in the backlog.  The board is sorted using the rank field.  So I have manually ordered the issues in the backlog then exported the issues that includes the rank field.

I have deleted the project, recreated it then done a system import using the export file created that includes the rank field.  I go into the board, the records are in the right order.  I run a filter showing the records sorted in order by rank.  All good, or so I thought....

This is where it gets strange: In the backlog, I move an issue to change its order.  I refresh the screen then all of a sudden, all of the issues change their order into what appears a random order.  I run a filter showing the records sorted in order by rank and they are out of order as well.  So by changing the order of one issue seems to have messed up the ordering.

There's something strange going on here and I'd appreciate some suggestions on why the order is changing and a way to resolve it.

Many Thanks
David

[Andy Heinzer]

@David Stokes What version of Jira is this?  It sounds like there is something strange happening in regards to the rank values being changed.  Off-hand it could be there is a rebalancing event happening on the system.   If you're a Jira System Admin, you can navigate to the Cog icon -> System -> Lexorank management or go to /secure/GHLexoRankManage.jspa

This page can tell you a lot more info about your ranking management.  There are also some integrity checks you can run at the bottom of the page.  I would be interested to see what you get back there. 

[William Crighton [CCC]]

After much experimentation with preserving rank using csv export/import we have determined the following - the sort order of your import file is extremely important. That is you need to sort your import file with the Rank column and you must not enable 'Natural sort' when doing so.

From what we can tell the Rank value is not being imported. This might be due to our source system not having a balanced Rank field, we are not sure. But from what we can tell the rank value is set according to the sort applied to the import file.

When we sort the import with issue key ASC the relative ranking is not preserved even when mapping the Rank value (we mapped the Rank value for every import - it never seemed to matter). When we sort the import with Rank and 'Natural sort' enabled the relative ranking was not preserved. 

It was only when we sorted the import csv file using Rank with 'Natural sort' NOT enabled that we were able to preserve the relative ranking in our source system.

Hope this helps someone, cheers.