MongocampCollectionCopy
Copies the documents of one collection into a new collection, inside the database (server 1.8.1) — $out and $merge stay blocked. It has a quick backup that picks a timestamped name for you. A dialog without a trigger: the parent owns v-model:open. MongocampCollections opens it from the Copy collection button (administrators, ordinary collections only).
Usage
vue
<script setup lang="ts">
const open = ref(false)
</script>
<template>
<UButton label="Copy" @click="open = true" />
<MongocampCollectionCopy v-model:open="open" source="orders" @done="reload" />
</template>Props and events
| Name | Type | Description |
|---|---|---|
v-model:open | boolean | Whether the dialog is open; the names and the document count are read fresh every time it opens |
source | string | The collection to copy |
filter | string | Optional Lucene filter: copy only the matching documents |
done | { target, copied } | Emitted after the new collection exists (also when the dialog was closed meanwhile), so a list can reload |
What it does
- Copy — type the name of the new collection (letters, digits and
_ - ., at most 100 characters; notmc_*,*.files,*.chunks,system.*, the source itself or any collection that exists, ignoring case; the form says which rule fails). - Quick backup — one click, no name: the new collection is called
<source>_backup_<yyyymmdd>_<hhmmss>(local time;_2,_3, … when that name exists). Odd characters in the source name become_and a long name is shortened. - One request. The server copies the documents (the count is shown beforehand; a filtered dialog copies only the matches). While it runs the dialog shows a progress bar without numbers and cannot be dismissed, because the work happens on the server.
- Also create the same indexes (a checkbox, on by default; it applies to the copy and to the quick backup): the server recreates every index except
_id_from the real index definitions — compound order, unique, expiry, partial filters, text weights, collations. The result says how many were created.
What is not copied
Grants, the schema and the files of a GridFS bucket are never copied — only documents and, if ticked, indexes.
Limits and failures
- No document cap (the old browser-side copy stopped at 50 000).
- If another session took the name meanwhile, the server answers
409; the form says so and asks for another name (a quick backup chooses a new name by itself, up to three tries). - Any other refusal shows the server's message and keeps the form.