How to Import a CSV into Power BI (3 Ways)

Getting a CSV into Power BI sounds trivial until the data lands in one giant column, your accents turn into question marks, or every number imports as text. The mechanics are simple, but the choices you make in the import dialog decide whether the rest of your build is smooth or a fight. This guide covers the three imports that actually matter — a single file, a whole folder, and pasting data by hand — plus the delimiter, encoding, and data-type traps that catch people.

All of this happens in Power BI Desktop, which is free to download and use. You don't need a Pro license or the Power BI Service just to import a file and build a report locally; a paid license only comes into play when you publish and share reports in the cloud.

Way 1: Get Data > Text/CSV (the standard single-file import)

This is the one you'll use most. For a single .csv (or .txt) file:

1. Open Power BI Desktop and click Home > Get Data.
2. Choose Text/CSV (it's near the top of the common sources).
3. Browse to your file and click Open.
4. A preview window appears showing a sample of the rows and three dropdowns at the top.

Those three dropdowns are the whole game:

• File Origin — the character encoding (more on this below).
• Delimiter — the character that separates columns.
• Data Type Detection — whether Power BI samples the first 200 rows, the whole file, or nothing to guess column types.

Check the preview. If the columns look right, you have two buttons: Load and Transform Data. Pick deliberately — the difference matters enough that it gets its own section below.

Way 2: From Folder (combine many CSVs at once)

If you get a new export every week — sales_jan.csv, sales_feb.csv, and so on — don't import them one at a time. Point Power BI at the folder instead.

1. Get Data > More > Folder (under the File category).
2. Browse to the folder and click OK. You'll see a list of every file in it.
3. Click Combine & Transform Data.
4. Power BI shows a sample file, confirms the delimiter and encoding once, then stacks all matching files into a single table.

This is one of the most underused features in Power BI for repetitive reporting, and the payoff is real:

• Every file with the same columns gets appended into one table automatically.
• A Source.Name column is added so you can tell which file each row came from — handy when a date or region is baked into the filename.
• When you drop a new file into the folder and hit Refresh, the new rows appear with zero extra work.

A few cautions:

• All files should share the same structure (same columns, same order). Mismatched headers cause errors or null columns.
• Filter out junk first. If the folder also contains a ~$ lock file, a README.txt, or subfolders, add a filter step on the Extension or Name column in Power Query before combining.
• Behind the scenes Power BI builds a small sample-file query and a transform function. Don't delete those helper queries — the combine breaks without them.

Way 3: Enter Data (paste a small table by hand)

Sometimes you don't have a file at all — you've got a lookup table in your head, a quick mapping, or a few rows copied from an email. Use Home > Enter Data.

A grid opens. Type or paste your rows, double-click the column headers to rename them, name the table, and click Load (or Edit to open it in Power Query first).

Good uses:

• Small static lookup tables — region-to-manager, status codes, category groupings.
• A manual date or parameter table while you prototype.
• A clean block of cells pasted straight from Excel.

Skip it for anything that changes. Enter Data hard-codes the values into the model, so there's no refresh — if the source updates, you have to edit the table by hand. For recurring data, save it as a CSV and use Way 1 or Way 2 instead.

Delimiters: when your data lands in one column

If your import shows everything jammed into a single column, the delimiter is wrong. The Delimiter dropdown in the Text/CSV dialog lets you choose:

Delimiter	Common source
Comma	Standard US/UK CSV exports
Semicolon	European exports (where comma is the decimal separator)
Tab	.tsv files and many database dumps
Colon / Space	Log files and some legacy systems
Custom	Pipe (|) and other characters — choose Custom and type it in

"CSV" technically means comma-separated, but in practice the separator varies by region and tool. If a colleague in Germany sends you a file that looks broken, try the semicolon first.

Encoding: when accents and symbols turn into garbage

If names, currency symbols, or accented characters import as é or ?, the File Origin (encoding) is wrong. Power BI guesses, and it doesn't always guess right.

• For modern files, choose 65001: Unicode (UTF-8). This is the safe default for almost everything today.
• Older Windows exports sometimes need 1252: Western European (Windows).
• A file that opens fine in Notepad but breaks in Power BI is almost always an encoding mismatch — change File Origin and re-check the preview.

You can't fix encoding after loading; you have to re-import with the right setting, so get it right in the dialog.

Transform vs. Load: which button to click

This is the single most important decision in the import.

• Load brings the data into the model exactly as previewed. Fast, and fine for clean files.
• Transform Data opens the Power Query Editor first, so you can shape the data before it enters the model.

Reach for Transform Data in almost every real-world case, because raw CSVs are rarely clean. In Power Query you can:

• Promote the first row to headers (or remove junk rows above the real header).
• Set data types explicitly — see the next section.
• Trim whitespace, replace blanks, and split or merge columns.
• Filter out totals rows, blank rows, or files you don't want.
• Unpivot wide, month-across-the-top data into a tall, model-friendly shape.

Every step you take is recorded in Applied Steps on the right. It re-runs automatically on each refresh, so cleaning a messy export once means it stays clean. That's the difference between a report you fight every week and one that just refreshes.

Don't trust auto-detected data types

Power BI samples your rows and assigns a type to each column. It's usually close, but the misses cause the worst downstream bugs.

The classic trap: a ZIP code, account ID, or product SKU like 00123 gets read as a number, drops its leading zeros, and silently corrupts your data. Or a column with a few stray text values forces the whole thing to Text, so your numbers won't sum.

Fix it in Power Query by clicking the small type icon to the left of each column header and choosing the right type:

• Whole Number / Decimal Number for amounts and counts.
• Text for IDs and codes that have leading zeros or aren't math.
• Date for dates — and watch for day/month order (03/04/2026) on international files. Use Change Type > Using Locale to force the right interpretation.

Rule of thumb: anything you will SUM or AVERAGE -> a number type.
Anything that is an identifier, code, or category -> Text.
Anything you will filter or trend over time -> Date.

Setting types correctly here also keeps your model lean and makes relationships and DAX behave. Once a column is a proper number type, a basic measure just works:

Total Sales = SUM ( Sales[Amount] )

If Amount had imported as text, that measure would error until you fixed the type — so the cleanup pays off immediately.

After the import: think in tables, not one big sheet

A single CSV loads as a single table, which is fine to start. But the moment you bring in a second related file — orders plus a customer list, say — resist the urge to merge everything into one flat table. Power BI is built for a star schema: slim fact tables related to lookup (dimension) tables. Use From Folder to stack files of the same shape, and relationships to connect files of different shapes. That keeps your model fast and your DAX simple. If you want to go deeper, see our guide on building a star schema in Power BI.

Once the data is in and typed correctly, you're ready to build — and if you want the report to look polished from the first visual, our free Power BI Theme Generator gives you a branded color theme to drop in.

---

Rather skip the import dialog entirely? That's what Instant PowerBI is for. Send us your raw CSV or Excel files and we'll hand back a modeled, branded Power BI report — delimiters, encoding, data types, relationships, and measures handled for you — that opens right in free Power BI Desktop. Get started and let us do the cleanup.