Importing JIRA project from XML generated in Issue Navigator

[Rutger S]

Hi,

We're about to import an Agile project from another company's JIRA instance. There are multiple projects in there so there's a huge chance we cannot receive the complete database export XML (generated by the usual exporter).

I thought of an alternative and the question is: is this possible?

• Select the project in the Issue Navigator.
• View the XML format of the project and save the XML file (export).
• Analyse the content of source and target XML formatting, and transform the source XML in such a way that all data match the target, i.e. that it can be imported regarding all settings in the target, in an empty Agile project. E.g. issue type, workflow status, etc. etc. should match (so for some non-existing in the target, other values will have to be chosen). This sounds like a delicate 'search and replace' mission. 
• Question: I read that active objects, e.g. Sprint, will not be imported. However the Sprint name is included in the XML. So...? If indeed the Sprints are imported, how will the become linked (part of) the to-be-made Agile board for the project? Sounds like this should be done some other way, or the Sprints are indeed simply not imported. So how does it work? When to create an Agile board and how does it link to the project? Should sprints be re-entered manually perhaps? I named 'sprint' as an example, but in the XML there are several more, mostly starting with 'com.pyxis.greenhopper.jira' - but also the (new) rank, as a custom field type.
• Question: can the non-used data be omitted? E.g. if neither their nor our JIRA uses Epic Color, can we lave it out of the XML?
• Copy the XML file to the import dir.
• Create users from the XML file if they didn't exist yet, and create an empty project for the project to be imported.
• Select the attachments from the source att dir and copy them (in their unique att subdirs) to the target import att dir, as specified in the JIRA importer.
• Question about certain data in XML that seem incomplete: e.g. <watches>2</watches> shows there are 2 watchers for this issue, but their usernames are not included. Is there anything to solve this or are these data simply missed after importing? In that case it seems <watches>2</watches> (and perhaps others) can be deleted from the XML anyway.
• Critical question for all globally used values, such as id's (project id, issue key id, rank, etc.): since the source id's in the XML file may already exist in the target JIRA, to what extent does JIRA check/re-number these automatically? Of course, nothing should be overwritten merely because an id accidentally already exists on the target. E.g. this goes for attachments too. Each att is located in its own dir, named after its unique id. So, if one source att is located in /123456/ while this dir/att already exists in our target JIRA, will JIRA check this automatically and assign higher numbers when importing? Of course, I hope so; how else should I think of a 'high enough' issue rank, for example, in its unreadable format? For the numerous unique id fields, this would require a lot of manual labour too it seems. (Strange I couldn't find anything about this topic, btw, or did I miss something?)
• Just for sure, create a backup of the target JIRA database.
• Use the JIRA importer for XML, specify the file and hope to have some luck.
  
  

Any other alternatives, hints, tips? Missed something else? Thanks.

[Boris Berenberg · Accepted solution]

JIRA XML importer will not handle the XML you are discussing. I would use the Excel output format, and then convert it to CSV, and then import using the CSV importer.

While it is possible that the Sprints may work by doing this, I expect that you will have to have already created identical boards and sprint names first.

[Rutger S]

Hi Boris, thanks. Why would the XML importer not handle the XML output from the issue navigator? Sounds a bit strange (but I'm definitely no expert in this field, there may be a good reason). How about the globally used values that may already exist? No overwriting if they do? This seems to go for the rank (lexo-rank) mainly. This may contain values such as 1|xzzwvc:. Strangely, the rank values themselves are empty in XML, but present in Excel. All other global ordering numbers (id's) seem not present in xls/CVS, as opposed to XML so they should pose not problem anyway.

[Rutger S]

It proved to be a bit complex after all, so I created a ticket at Support. Thanks so far anyway! For other readers trying to import Agile data, please refer to https://jira.atlassian.com/browse/JRA-28748