Launch PowerShell the same way, using powershell instead of cmd, and run:

Get-ChildItem *.csv | ForEach-Object { Import-Csv $_.FullName -Encoding UTF8 } | Export-Csv merged.txt -NoTypeInformation -Encoding UTF8

In our Windows 11 test, piping Get-ChildItem straight into Import-Csv under 5.1 failed with CannotSpecifyPathAndLiteralPath. ForEach-Object avoids the error by passing one path at a time.
Without the switches, the result began with #TYPE System.Management.Automation.PSCustomObject, and the customer name Zoë became Zo?. Microsoft's Export-Csv reference for PowerShell 5.1 confirms: "The default value is ASCII." PowerShell 7 writes UTF-8 without a BOM and skips the #TYPE line by default.

1 of 5 columns

Microsoft states: "If the remaining objects have additional properties, those property values are not included in the file." With -Append, mismatched columns stop the command with an error. Adding -Force writes the file but discards the mismatched properties.
Import-Csv reads every value as text; Export-Csv quotes each value by default. A user on PowerShell.org found that values "turned into string" after merging two files. Excel converts them back to numbers when opening the file, but a database import may not.

Put the CSV files in a separate folder with nothing else in it.

• In Excel, click Data > Get Data > From File > From Folder, pick the folder and click Open
• In the file list, click Combine > Combine & Transform Data
• In the Combine Files dialog, check File Origin and Delimiter on the sample file, then click OK
• In the Power Query Editor, check each column type and click Close & Load

As The Bricks notes, "Power Query tries to guess the data type for each column." It uses the first rows, so ZIP codes can become numbers and lose their leading zero. Set those columns to Text in the editor.
The query stays linked to the folder. Add next month's export and click Data > Refresh All to update the table. We'd choose this route for a monthly report using the same export each time.

One sheet, or one sheet per file?

For separate worksheets, import each file with Data > From Text/CSV and load it to a new sheet. For .xlsx workbooks, use Move or Copy as described in merge two Excel spreadsheets.

Why does Power Query drop columns from some files?

Clicking Combine builds a helper query from the sample file, usually the first file in the folder. Its source step uses a fixed setting such as Columns=4. The accepted answer in a Tech Community thread about lost columns says: "Change Columns=10 on Columns=20 or like, i.e. max possible number of columns."
Open the Transform Sample File query, select Source and edit the number in the formula bar. If the new column still doesn't appear, check Expanded Table Column1 in the main query. It lists column names from the sample file and leaves out names missing from that list.

What if the CSV files have different columns?

$rows = Get-ChildItem *.csv | ForEach-Object { Import-Csv $_.FullName -Encoding UTF8 }
$cols = $rows | ForEach-Object { $_.PSObject.Properties.Name } | Select-Object -Unique
$rows | Select-Object $cols | Export-Csv merged.txt -NoTypeInformation -Encoding UTF8
This script kept all five columns in our three test files, in the order first seen. January and February had empty Region cells; March kept its regions.
The pandas.concat documentation describes the default outer join: "Columns outside the intersection will be filled with NaN values."

import glob, pandas as pdframes = [pd.read_csv(f, encoding="utf-8-sig") for f in sorted(glob.glob("*.csv"))]
pd.concat(frames, ignore_index=True).to_csv("merged.csv", index=False, encoding="utf-8-sig")
Header names must match exactly, apart from case in some programs. "Email" and "E-mail" become separate columns in every method here, ours included. Rename conflicting headers before merging. For example, a phone export from a VCF to CSV converter and a CRM export may name the same field differently.

If exports contain the same orders or customers, compare two Excel sheets before stacking them to find repeated rows. Remove duplicates once after merging.

Why do merged CSV files show garbled characters?

If the bytes are correct but Excel reads them with the wrong encoding, re-importing with the right File Origin fixes the display without changing the file. A Discourse Meta discussion explains that Excel needs encoding metadata, which a BOM provides.
If characters were lost when writing, as in our PowerShell 5.1 test, no import setting can recover them. Re-run the merge from the original files with explicit UTF-8 encoding.

The copy command carries each source file's BOM into the merged file at the start of its header line. Deleting the extra header lines removes those BOMs too.

When the delimiter differs between files:

European-language versions of Excel and accounting programs often export semicolons because they use commas as decimal marks. If you merge those files with comma files, each semicolon row lands in one column. The e-iceblue CSV merging guide advises: "If your files use a different encoding or delimiter, adjust the import settings accordingly before processing." In PowerShell, read those files separately with Import-Csv -Delimiter ';'.

Excel and Import-Csv handle quoted fields containing line breaks. A line-based trick that deletes "line 1 of each file" can split those rows.

What happens when the merged file passes 1,048,576 rows?

1,048,576 rows

Source

16,384 columns

Source

Twelve monthly exports of 90,000 rows make 1,080,000 rows. Open that file directly and Excel loads only what fits, with the warning quoted in Microsoft's support article on oversized data sets: "This data set is too large for the Excel grid. If you save this workbook, you'll lose data that wasn't loaded."

For analysis, merge with Power Query and choose Close & Load To > Only Create Connection, with Add this data to the Data Model checked. Memory limits the Data Model: Microsoft's specifications give 64-bit Excel no hard file size limit and 32-bit Excel a 2 GB address space. For another system, keep the merged result as CSV and skip Excel.

You don't need to install anything. ExtendsClass offers "Keep only the header (first row) of the first file" and states on its CSV merge page: "Your files are not sent to the server side, your CSV files are merged directly in your browser!" That is the vendor's claim as of October 2026. We did not test online programs for this guide.

You can check in two minutes: open the program, press F12, switch to the Network tab and merge a dummy file. An upload request roughly the file's size means it left your PC. No request means the merge ran locally.
A browser merger holds every file in the tab's memory, so large merges can slow down or fail. Its handling of mixed headers and encodings also depends on the program.

When does a desktop CSV merger beat the free methods?

You get every header in the order first seen, with empty cells where a file lacks a column. Matching ignores letter case and spaces at the edges of headers. For a file starting with a title or logo row, set the header row number to skip the rows above it.
The program detects each file's delimiter (comma, semicolon, tab or pipe) and encoding (BOM, UTF-8, ANSI or Latin-1). A quoted multi-line comment stays one row. Damaged files are skipped and counted in the final message. At Excel's row limit, the program stops, keeps what it has merged and tells you to split the batch. You can merge CSV and Excel files into one workbook, saving as .csv or .xlsx on one sheet or one sheet per file.

Resave old .xls files as .xlsx before adding them. Matching uses exact header text after trimming, with no synonyms. The program runs only on Windows and has no command line for scheduled runs. The trial writes the first 50 data rows, excluding the header; a license removes that limit.

Sources

• Export-Csv reference for PowerShell 7 (Microsoft Learn) utf8NoBOM default, -Append and -Force behavior with mismatched columns
• Loading a large text file into the Data Model (Microsoft Tech Community) Data Model size limited by PC memory, not the grid