How to completely remove documents from collection

[daniel-kennedy]

I have a collection "Departments" whose documents are in a JSON file. When I add a document it appears in results immediately; if I remove a document, however, it still appears in API results and in the mini-dashboard, so I am assuming it is cached somewhere. How do I sync the cache to the JSON file, so that they are in agreement?

Thanks!

[curquiza]

Hello @daniel-kennedy

Deletion of document is a task that can take time to be processed by meilisearch. Maybe the deletion was not completed yet?

Also, can you check you have the similar issue by directly calling the /search route using curl command and not the front-end dashboard?

[daniel-kennedy]

Thank you for responding. I used curl and postman and the documents I removed from the JSON file are still being returned as results. It has been over a week since I made the changes. Perhaps my process is not correct? I edit the index object, in my case a file named departments.json, and then load the index via a python script.

import meilisearch
import json

client = meilisearch.Client('http://localhost:7700', '**myAdminKey**')
json_file = open('departments.json', encoding='utf-8-sig')
departments = json.load(json_file)
client.index('Departments').add_documents(departments)

Visiting /tasks returns tasks that all have enqueued/started/finished times to date. I can see where task 59 is the one where I attempted to remove three and although it shows it as succeeding (the number of documents drops), the prior results still come back.

Thanks again,
Dan

[daniel-kennedy]

I got it removed by http verb. It makes sense that it behaves like an API and that I can manage it as such.

My apologies!

[curquiza]

Hello @daniel-kennedy
No worries, indeed use the method delete_documents to remove them, which use the DELETE HTTP verb. Otherwise they will be kept in the engine 😊