The route

Your first pipeline: CSV to console

Download the exercise files

Two nodes and one run. By the end of this you'll have read a real CSV file into Odara, printed it, and — more useful than either — learned where to look to find out what actually happened.

Goal

A pipeline with a CSV Source feeding a Console node, run once, printing the first ten rows of customers.csv.

You'll need

  • Odara installed and open (World 1 · Install and first launch).
  • The exercise files for this level — the Download exercise files button above. Unzip them into a folder you'll keep for the whole course; the screenshots use /tmp/odara-academy/, yours will differ.
  • About 15 minutes.

The archive holds one file, customers.csv: 500 rows, 9 columns, comma-delimited, UTF-8, with a header row. It's the customer list of a fictional SaaS, and it comes back in later levels — World 3 joins it against orders, World 4 loads it into Postgres.

Steps

1. Create the pipeline

In the Explorer tab on the left, hover MY PIPELINES and click + (or right-click it and choose New Pipeline). A pipeline called New Pipeline appears with its name ready to edit — type w1-3-first-pipeline and press Enter. It opens as an empty canvas in its own tab. There's no save button to remember: the top bar shows Saved as you work.

A new, empty pipeline with its name being edited

2. Drop a CSV Source on the canvas

Switch to the Nodes tab. The palette is split into four sections — INGEST, TRANSFORM, DELIVER and ORCHESTRATE — and each section into folders. Open INGEST ▸ Files and drag a CSV Source onto the canvas.

Click it. The Properties panel on the right opens with the node's name at the top; use the pencil next to it (Rename) to call it customers. Names show up in the log and in the generated docs, so name nodes for what they hold, not what they are.

Then fill in its configuration:

Field Value
File Path the absolute path to your customers.csv
First row is header checked
Delimiter ,
Row Limit leave empty — it reads every row

The customers node selected, with its Properties panel

File Path is read by the Odara server, not by your browser. When Odara runs on your own machine that's the same thing; when it runs on a server or in a container, the path has to exist there.

3. Preview before you run

At the bottom of the Properties panel, click Preview Output. Odara reads the file and shows you the first rows along with the total: Showing 10 of 500 rows.

Preview Output: the first rows and the total row count

Make this a habit. Preview answers "did I point it at the right file, with the right delimiter and header setting?" in two seconds, before anything downstream depends on it.

4. Add a Console node and connect it

From DELIVER ▸ Debug, drag a Console node to the right of the source and name it console. Drag from the source's output handle (right edge) to the console's input handle.

Console prints rows into the run log instead of writing them anywhere. It's the debugging node: when a pipeline misbehaves, dropping one of these at any point shows you what the data looks like there.

Its Row Limit defaults to 100. Set it to 10 — that's plenty to eyeball.

5. Execute and read the log

Click Execute in the top bar. A Completed in …ms badge appears over the canvas.

Now open Console in the top-right corner. It's the run log, and it tells you, per node, what happened: customers: 500 rows, then console: 500 rows, and a [console] Output (click to expand) entry. Click that entry to see the table.

The Console panel with the printed table and the per-node row count

Expected result

  • The run completes, with no failed node.
  • The log shows customers: 500 rows.
  • The expanded Console output is a table of 10 rows, the first one being customer 1001, Ursula Cardoso, from IT, on the free plan.
  • Nine columns, in file order: customer_id … active.

Common pitfalls

Preview fails and the message mentions the path. The file isn't where the server is looking. On Windows, a path copied from Explorer sometimes arrives wrapped in quotes — remove them. Use the absolute path.

Every row comes out as a single column. The delimiter doesn't match the file. Open the CSV in a text editor — not Excel, which hides exactly the character you're trying to see — and look at the first line. World 2 hands you a file where this happens on purpose.

The columns are called column_1, column_2, … and there are 501 rows. First row is header is unchecked, so the header line became a data row. Check it and preview again.

The run is green but the console table is empty. The two nodes aren't actually connected — look for the line between them on the canvas. A node with no input runs and does nothing, and Odara doesn't treat that as an error.

What you built

A source, a target and a connection: that is the whole shape of every pipeline you'll build from here. World 1 finishes by putting a transform between them and writing the result to a file instead of the screen.

Check yourself

One question, drawn from a larger bank. Get it right to pass.

  1. You want the console to print 25 rows instead of 10. What do you change?