nlqdb

Solve · Analysts and PMs

How do I combine multiple rows into one comma-separated value in SQL?

If you want one row per customer with their products rolled into a single comma-separated list — not one row per order line — ask in plain English. nlqdb compiles the string aggregate (`STRING_AGG` in Postgres, `GROUP_CONCAT` in MySQL), orders the items inside the aggregate so the list is stable, and shows the SQL.

Rolling many rows up into one value — every product a customer bought as a single 'apples, bananas, cherries' cell, all tags on a post in one line, every error code per host — is the report shape a plain `GROUP BY` can't give you: grouping collapses the rows but leaves you needing an aggregate that concatenates text, not one that sums numbers. Every engine spells it differently — `STRING_AGG` in Postgres and SQL Server, `GROUP_CONCAT` in MySQL, `LISTAGG` in Oracle — and two traps bite: without an explicit `ORDER BY` inside the aggregate the list order is undefined and shuffles between runs, and the delimiter is yours to choose, so a value that itself contains a comma can make the joined string ambiguous.

The snippet that solves it.

> each customer's product names combined into one comma-separated list, alphabetized

What nlqdb does for this

  • Ask 'each customer's product names combined into one comma-separated list'; nlqdb compiles the `STRING_AGG` string aggregate and runs it in Postgres.
  • It groups by the customer and rolls each group's rows into one delimited string — one row per customer, not per order line.
  • It orders the items inside the aggregate, so the list is stable across runs instead of shuffling in whatever order rows happened to arrive.
  • Every answer shows the compiled SQL under a trace toggle, so you can confirm the delimiter and the in-aggregate ordering are what you meant.

Drop into any HTML page

<nlq-data goal="each customer's product names combined into one comma-separated list, alphabetized"></nlq-data>

The roll-up you'd otherwise hand-write — the string aggregate, the delimiter, and the ORDER BY inside it that keeps the list stable — is one English goal here, with the SQL shown so you can confirm the items come back in the order you asked for.

What this replaces

What nlqdb doesn't try to do here

  • The delimiter and the item order are ones you name — 'comma-separated, alphabetized'. nlqdb compiles and shows them; it won't guess a separator or a sort for you.
  • This is the roll-up direction only — combining rows into one value. Splitting a delimited string back into rows (the inverse) is a different query, not this page.
  • The public `<nlq-data>` embed is read-scoped — it returns the joined list, it doesn't write. Persisting the rolled-up value goes through the SDK or `POST /v1/run`.

Questions buyers ask

How do I combine multiple rows into one comma-separated value in SQL?
Ask in plain English — 'each customer's product names combined into one comma-separated list.' nlqdb compiles a string aggregate — `STRING_AGG(product_name, ', ' ORDER BY product_name)` in Postgres — groups by the customer, and orders the items inside the aggregate so the list is stable. It runs the query in Postgres and shows the SQL, so you can confirm the delimiter and the ordering.
What's the difference between STRING_AGG and GROUP_CONCAT?
They're the same idea in different dialects: `STRING_AGG(expr, delimiter)` in Postgres and SQL Server, `GROUP_CONCAT(expr SEPARATOR ',')` in MySQL, `LISTAGG(expr, ',')` in Oracle. All roll grouped rows into one delimited string. nlqdb targets Postgres, so it compiles `STRING_AGG` and shows it — you don't have to remember which engine spells it which way.
Why does my concatenated list come back in a random order?
Because a string aggregate has no inherent order — without an explicit `ORDER BY` inside the aggregate, the database concatenates rows in whatever order it read them, which can change between runs. Put the sort inside the aggregate (`STRING_AGG(name, ', ' ORDER BY name)`), not in the outer query. nlqdb compiles the in-aggregate ordering and shows it in the SQL.
Can I roll up rows on a Postgres database I already run?
Yes — connect it with the signed-in BYO connect verb (`nlq db connect`; see /solve/query-existing-postgres-in-natural-language) and ask for the combined list in place, no ETL into a separate store. The honest limits: BYO connect is signed-in only (not the public embed), and the answer is read-only — nlqdb returns the joined string, it doesn't write it back to a column.

Where this pain shows up in public

Enduring discussion hubs where you can verify the theme without taking our word for it. We don't quote individual posts; we cite search-result and subreddit URLs that stay live as new threads land.

Related guides

Try nlqdb in 30 seconds

No sign-in. The anonymous database lasts 72 hours; adopt it with one click if you keep it.

Start with a goal →

Looking at this from a different angle? Browse all solve pages or browse competitor comparisons.