Task Data Tables - TATER Help

What it is

A data table is a small grid attached to an Ops task. Where a comment is narrative ("checked srv1, looks fine"), a table row is a datapoint: a reading, a count, a pass or fail per device. Rows are timestamped and attributed to whoever recorded them, and they never appear in the task's comment thread or activity stream.

A task can carry several tables, each with its own columns. A task with a Disk free by server table and a separate Backup result by night table keeps the two apart, and each exports on its own.

Creating a table

Open a task in TATER Ops and find the Data tables panel under the worklog. Choose + New table, name it, and define the columns. Each column has a name and a type:

text - free text, up to 2,000 characters
number - anything that parses as a number, so it can be summed and charted once exported
date - a calendar date such as 2026-10-02, or an ISO timestamp
boolean - yes or no
select - one value from a list you provide, which keeps spellings consistent

Limits: 20 tables per task, 25 columns per table and 5,000 rows per table. A table that fills up should be exported and replaced by a new one.

Recording check-ins

Expand a table and use the entry row at the bottom. + Add records a row, stamped with who added it and when. The pencil on an existing row loads it into the entry row so you can correct it; only the cells you change are written.

A bad value is refused with a message naming the column, so a mistyped number never lands as text.

Changing columns later

Columns on a table lets you rename it, relabel a column, add columns and remove columns.

Each column has a permanent internal key, so relabelling a column never loses what was recorded under it.
A column's type cannot change once it exists. Add a new column instead.
Removing a column deletes its recorded values, and TATER asks you to confirm first.
A select option that existing rows still use cannot be removed.

Exporting

Each table has CSV and XLSX buttons. When a task has more than one table, All tables (XLSX) produces one workbook with a sheet per table.

Every export includes Recorded at and Recorded by columns ahead of your own.
Numbers and yes/no values stay typed in XLSX, so Excel can total them.
Text that begins with =, +, - or @ is prefixed with an apostrophe so a spreadsheet cannot run it as a formula.
CSV covers one table, because a CSV file cannot hold several.

Who can do what

Auditor and above can read tables and add, edit and remove rows.
The table's creator, or an Admin, can delete a whole table.
A task in a restricted queue shows its tables only to people who can see the task.

Through the AI Analyst and other agents

Agents write to the same tables, and several can check in at once without overwriting each other.

list_task_data_tables - the tables on a task, with their columns; pass a table id to read its rows
create_task_data_table - define a table, optionally seeded with first rows
add_task_data_table_rows - record check-ins. Rows can name a column by label or key, and a bad value rejects the whole batch
update_task_data_table_row and delete_task_data_table_row - correct or remove a row
update_task_data_table and delete_task_data_table - change the columns, or remove the table

Ask the Analyst to "log today's reading on this task" and it adds a row; it does not write a comment.

TATER Ops overview - where tasks live
On-Call Rotations