Sum / Average Records

Count, sum, average, min or max a field over matching records, optionally per group with a top list: revenue, best sellers (Bottly storage or your PostgreSQL).

Database — database

Bottly storage keeps records on Bottly (plan row limits, Storage page with import/export). Your PostgreSQL writes to your own database instead: the table and column names are used exactly as written, and values are sent as text parameters.

Input: Fixed choice.

Editor guidance: Bottly storage is built in (plan row limits apply). Your PostgreSQL uses your own database: tables and columns are used by name, values are sent as text parameters.

Available choices (preserve the exact value):

Connection string (Your PostgreSQL) — connection

A postgres://user:password@host:5432/database URL, kept in a bot env secret such as {env.DATABASE_URL} rather than typed into the block. The host must be reachable on the public internet; TLS is used unless the URL ends with ?sslmode=disable (?sslmode=verify-full also checks the certificate).

Input: Text with variable suggestions.

Editor guidance: Only used with Your PostgreSQL. Empty uses {env.DATABASE_URL}. Keep it in an env secret, never typed here. The host must be public; TLS is used unless the URL says sslmode=disable.

Format example, not a ready value for your server: {env.DATABASE_URL}.

Table — table

Name the bot-owned record collection consistently, such as tickets. A typo addresses a different table, not an alias of the old one.

Input: Text.

Format example, not a ready value for your server: orders.

Default: orders.

Filters (one per line) — filters

One condition per line: field = value, or != > < >= <= and contains (case-insensitive). Every line must match; leave it empty for all records. Values that are both numbers compare as numbers (10 > 9), anything else as text, and a missing field counts as empty, so sold != yes also matches records that were never marked. Values may use variables such as {user.id}.

Input: Multiline text.

Format example, not a ready value for your server: status = paid.

Default: ``.

Operation — operation

Choose the specific operation first, then fill only the inputs it requires. Different operations can return different data and can have different side effects.

Input: Fixed choice.

Default: sum.

Available choices (preserve the exact value):

Field (numbers) — field

The stored field name to compare. Match the name used when creating the record.

Input: Text.

Format example, not a ready value for your server: price.

Default: price.

Group by field (optional) — groupBy

Optional field to group by: one result per value of this field, highest first, written to {agg.top} and {agg.0.key}/{agg.0.value}. Empty gives one total in {agg.result}.

Input: Text.

Format example, not a ready value for your server: product.

Default: ``.

Top groups (1-25) — top

How many groups to keep when grouping (1-25), highest result first.

Input: Number.

Default: 10.

Store as — storeAs

Choose a short result prefix such as result. Later blocks must use this same prefix and run after this block. It is not a Discord channel or a persistent table.

Input: Text.

Editor guidance: {agg.result}; with group by also {agg.top} (one "value: result" line each), {agg.0.key}, {agg.0.value}

Default: agg.

Read the result after this block, using its configured prefix and the fields supported by the selected operation. Do not assume every response has a text or id field.

Test a missing record, then save and read the same lookup value. Use a narrow filter for Update/Delete. Renaming a table or field changes where the lookup runs.

Save, validate and publish. Perform one real test and read the first relevant log error. [Testing](/docs/flow-testing) · [Variables](/docs/flow-variables) · [Permissions](/docs/flow-permissions).