Skip to content

Batch Operations

KoukeNeko edited this page Sep 9, 2026 · 4 revisions

Batch Operations

English · 繁體中文

Bulk creation

printf '%s\n' "First issue" "Second issue" > issues.txt
taiga batch create issue issues.txt --sprint sprint-27 --dry-run --json
taiga batch create issue issues.txt --sprint sprint-27 --yes --json

cat stories.txt | taiga batch create story - --status New --yes
taiga batch create task tasks.txt --story 51 --yes
taiga batch create epic epics.txt --yes

Each non-blank line in the input file is one subject, up to 1000 items and 4 MiB. Pass - to read from stdin.

One --status can be applied to the whole batch. Issues accept a shared --sprint; tasks must have a sprint, given either directly with --sprint or derived from a parent story with --story.

Limitations

Taiga's native bulk endpoints do not support a different description or assignee per item, and they do not guarantee an atomic transaction across the batch. When you need per-item fields, use the single-item create commands.

After a successful response the CLI checks the number of records returned. If it does not match what was sent, it reports ambiguous_commit and asks you to list and confirm the real outcome, so a blind resend cannot create duplicates. Non-interactive runs require --yes.

Bulk move and reorder

taiga batch move story --id 101,102 --sprint sprint-28 --yes
taiga batch reorder story --view kanban --id 102,101 --status "In progress" --swimlane Backend --yes
taiga batch reorder task --view taskboard --order 501=1 --order 502=2 --status "In progress" --yes

batch move and batch reorder support stories, tasks, and issues through Taiga's native bulk endpoints.

reorder needs a --view such as kanban, taskboard, or backlog, because ordering is relative to a particular board view. Give the complete order with --id, or position individual items with --order <id>=<position>.

Clone this wiki locally