Implement an ID generator?

In a few use cases, internal stakeholders have requested to create a human-friendly identifier with an incrementing sequence. For example, if creating an inventory of use cases, they would like to see “UC-00157” rather than “45fac3af-1bf3-4805-bf5a-9458f9a8a73e”.
This seems relatively straightforward to implement in a workflow, but where to store the stateful sequence? At the moment, I feel my best bet is to store this information as part of a workflow configuration. So one workflow would be called “Stateful configuration” and would store all the info needed for all the other workflows.
The best part is I can redeploy it without overwriting the existing configurations.

Any thoughts or ideas?

def variable = 'useCaseSeq'
def cfg = workflowDefinitionApi.getWorkflowDefinitionByProcessId('ucbConfiguration')
def seq = cfg.configurationVariables.get(variable).toInteger() + 1
def code = "UC-" + seq.toString().padLeft(5, '0')

workflowDefinitionApi.changeWorkflowDefinition(
	ChangeWorkflowDefinitionRequest.builder()
	.id(cfg.id)
	.configurationVariables([variable: seq.toString()]).build()
	)

arthurburkhardt

Posted 5 years ago · Edited 1 year ago·Last reply 3 years ago

17 comments

arthurburkhardt

OP3 years ago · Edited

hum… have you checked if the script task are correctly configured as groovy and not javascript?

Tyler Bass

·3 years ago · Edited

Yes, the script task is is correctly configured as groovy.

Given your use of the word “are,” you are making me feel inclined to try to split my script task into two: One to define the assetName and another to add the asset and the attributes.

arthurburkhardt

OP3 years ago · Edited

You’re supposed to use the issueId as the name of your issue, otherwise it just doesn’t work. => You can adapt the logic to whatever else you want.
Also, I recommend against using the issueApi => There’s a lot of hidden hardcoded logic. For example, you can’t report on issues (in hidden issue domains) in the usage analytics dashboards.

shubhangishekhar

·3 years ago · Edited

Hi @arthur.burkhardt
I have utilized the same code you shared for generating id whenever a new issue gets created . I am seeing same Issue ID generated with each time a new issue gets created. So seek your help to avoid duplicate IssueID creation. For reference attaching the code here:

def requesterId = userApi.getUserByUsername(requester).getId()
def descriptionString = execution.getVariable(“description”)?.toString() ?: “”
def communityId = responsibleCommunity
def relatedAssets = execution.getVariable(“relatedAssets”) ?: []

def addAttributeToAsset(assetUuid,attributeValue,attributeTypeUuid) {
if (attributeValue == null){
return;
}
attributeApi.addAttribute(AddAttributeRequest.builder()
.assetId(assetUuid)
.typeId(string2Uuid(attributeTypeUuid))
.value(attributeValue.toString())
.build())
}

def relatedAssetsList = []
relatedAssets.each{ relatedAssetId ->
def relatedAssetRef = RelatedAssetReference.builder()
.assetId(relatedAssetId)
.direction(true)
.relationTypeId(string2Uuid(impactsRelationId))
.build()

relatedAssetsList.add(relatedAssetRef)

}

def namePrefix = “ISSUE-”
def seq = 1 + (
assetApi.findAssets(FindAssetsRequest.builder()
.name(namePrefix).nameMatchMode(MatchMode.START)
.sortField(SortField.NAME).sortOrder(SortOrder.DESC).limit(10).build()
).results
.collect{(it.name =~ /${namePrefix}(\d{5})/).collect{it}.flatten()}
.collect{whole, seq -> seq.toInteger()}.find()?:0
)
loggerApi.info(“THE SEQUENCE NUMBER IS: ${seq}”)
def issueId = namePrefix + seq.toString().padLeft(5, ‘0’)
loggerApi.info(“THE ISSUE GENERATED IS: ${issueId}”)

def newIssueUuid = issueApi.addIssue(AddIssueRequest.builder()
.name(subject)
.description(descriptionString)
.priority(priority)
.responsibleCommunityId(communityId)
.relatedAssets(relatedAssetsList)
.categoryIds(classifications)
.typeId(string2Uuid(dataIssueId))
.requesterId(requesterId)
.build()
).getId()
addAttributeToAsset(newIssueUuid,issueId,issueIdTypeId)
addAttributeToAsset(newIssueUuid,issueClassificationCategory,issueClassificationCategoryAttributeTypeId)
execution.setVariable(“outputCreatedTermId”,uuid2String(newIssueUuid))

shubhangishekhar

·3 years ago · Edited

Hi Team

In my scenario I have to generate data issue asset in sequence. I am able to generate the sequence number but the same Id gets created again (e.g: All issues have issue id as ISSUE-00001). How to resolve it and make each sequence unique.

arthurburkhardt

OP3 years ago · Edited

Hey @shubhangi.shekhar.tietoevry.com,
You should try to pinpoint your issue and share the specific code if you hope anyone can help.

arthurburkhardt

OP4 years ago · Edited

For reference, I gave up the initial ID generator implementation because changing configuration variables was updating workflows every time.
Instead, I ended up just querying last results and incrementing the highest asset number.

pros: Performance, easy to code, easy to maintain, low risk of errors
cons: risk of reusing ID (in case asset is deleted)

The following code does:

lookup 10 assets matching the prefix (+ other constraints), ordered by name descending
extract the number from the asset name
get the highest value
if there are no assets, return 0
increment number by 1
format asset name

def namePrefix = "ABC-"

def seq = 1 + (
  assetApi.findAssets(FindAssetsRequest.builder()
    .name(namePrefix).nameMatchMode(MatchMode.START)
    .sortField(SortField.NAME).sortOrder(SortOrder.DESC).limit(10).build()
  ).results
  .collect{(it.name =~ /${namePrefix}(\d{5})/).collect{it}.flatten()}
  .collect{whole, seq -> seq.toInteger()}.find()?:0
)

def name = namePrefix + seq.toString().padLeft(5, '0')

The reason why 10 assets are looked up is to avoid the risk that incorrectly named assets could overshadow the last element of the sequence. (e.g. ABC-Z0001 would cause errors.)
If there is no risk for incorrectly named assets, you could adjust the limit(10) to limit(1) or whatever you like.

arthurburkhardt

OP4 years ago · Edited

pros: Performance, easy to code, easy to maintain, low risk of errors
cons: risk of reusing ID (in case asset is deleted)

The following code does:

lookup 10 assets matching the prefix (+ other constraints), ordered by name descending
extract the number from the asset name
get the highest value
if there are no assets, return 0
increment number by 1
format asset name

def namePrefix = "ABC-"

def seq = 1 + (
  assetApi.findAssets(FindAssetsRequest.builder()
    .name(namePrefix).nameMatchMode(MatchMode.START)
    .sortField(SortField.NAME).sortOrder(SortOrder.DESC).limit(10).build()
  ).results
  .collect{(it.name =~ /${namePrefix}(\d{5})/).collect{it}.flatten()}
  .collect{whole, seq -> seq.toInteger()}.find()?:0
)

def name = namePrefix + seq.toString().padLeft(5, '0')

Tyler Bass

·3 years ago · Edited

I am trying to implement an ID generator that only takes the highest primary-key value (just a number like “328”) in a domain and adds one to it. The console is providing somevague, negative feedback when trying to upload a workflow with the code following this paragraph. If you would have any constructive critiques, please, I would appreciate that. Thanks.

def assets = assetApi.findAssets(FindAssetsRequest.builder()
    .sortField(FindAssetsRequest.SortField.NAME)
    .sortOrder(SortOrder.DESC)
    .limit(1)
    .build()
)

def highestNumber = assets.results.first()?.name?.toInteger() ?: 0

def newId = highestNumber + 1

def assetName = newId.toString()

arthurburkhardt

OP3 years ago · Edited

Did you import all the required classes (FindAssetsRequest, SortOrder, etc.)?
If you can paste the logs, it would be useful to troubleshoot.

stefan_busch

·3 years ago · Edited

wouldn’t you need to call assets.getResults() to get the results (instead of assets.result)

arthurburkhardt

OP3 years ago · Edited

No, what @tyler.bass wrote is perfectly valid, this is the groovy shorthand and I prefer that syntax as well, that’s what I always use.

assets.getResults() == assets.results

This is applicable to all the getter methods

asset.getId() == asset.id
asset.getName() == asset.name
relation.getSource() == relation.source
...

stefan_busch

·3 years ago · Edited

thanks for the info Arthur!

Tyler Bass

·3 years ago · Edited

This is what has been coming out of the console:

org.hibernate.LazyInitializationException: failed to lazily initialize a collection of role: com.collibra.dgc.core.model.user.impl.RoleImpl.rightsSet, could not initialize proxy - no Session

Ahead of my script attempting to add assets and attributes, along with the ID-generation functionality, I’ve been importing like so:

import com.collibra.dgc.core.api.dto.instance.asset.AddAssetRequest;
import com.collibra.dgc.core.api.dto.instance.attribute.AddAttributeRequest;
import com.collibra.dgc.core.api.component.instance.AssetApi;
import com.collibra.dgc.core.api.component.instance.asset.FindAssetsRequest;
import com.collibra.dgc.core.api.dto.SortOrder;

Again, thanks for your help.

Ronald Schaareman

·5 years ago · Edited

Hi @arthur.burkhardt,
We had the same kind of request actually and Collibra built a custom workflow for us. In our case it was about change request (CR) numbering and the CR number (#) would be assigned as the Asset Name (format CR#) .

The workflow would substring all the CR numbers in the assigned domain, put it in a list, find the max number and create the next number (# + 1).

Maybe this gives you an idea.

Kind regards, Ronald

bartvanderlocht

·5 years ago · Edited

In the past, we implemented something similar with a ‘hidden domain’:

-create hidden domain (i.e. a domain only visible for sysadmins)
-create one asset per ‘counter’ you need to have
-modify (increment) the name of the counter-assets

Additionally, we had one ‘batch-workflow’ which could generate multiple keys at once. This is helpful when assets are also created outside of workflows (e.g. by excel-imports)

adityapawar

·5 years ago · Edited

Hi Arthur,

Previously when I have asked to build such requirement at that time I have implemented it by incrementing count of assets as mostly assets of that type are represented by different unique asset type and we never give option to delete assets to any user so you are just adding new assets that used to work for me.