readxl skip vs range in R: Which One to Use

In readxl, skip drops a fixed number of rows from the top of a sheet before reading, while range reads an exact rectangular block of cells. Use skip for stray title rows, and use range when you need precise control over both rows and columns.

read_excel("f.xlsx", skip = 3)                       # drop the first 3 rows
read_excel("f.xlsx", skip = 3, n_max = 100)          # skip, then cap rows
read_excel("f.xlsx", range = "B3:D90")               # exact A1 cell block
read_excel("f.xlsx", range = "Data!B3:D90")          # range with a sheet name
read_excel("f.xlsx", range = cell_rows(3:90))        # rows only, all columns
read_excel("f.xlsx", range = cell_cols("B:D"))       # columns only, all rows
read_excel("f.xlsx", range = anchored("B4", c(10,3)))# anchored 10x3 block

skip and range in one sentence

skip and range both control which part of a sheet read_excel() imports, but they work at different scales. The skip argument counts rows from the top and discards them. The range argument names an absolute rectangle of cells, so it controls rows and columns at once.

To see the difference, build a messy spreadsheet with two title rows above the real data. The writexl package writes the file so you can read it straight back.

Read the file with no arguments and the two title rows leak in as data, while the first title becomes a column name.

Using skip to drop rows from the top

The skip argument removes a set number of rows before readxl looks for a header. Pass skip = 2 to discard both title rows. The next row (product, units, revenue) is then treated as the header, and the three fruit rows become the data.

skip only affects rows. It keeps every column in the sheet, so it is the right tool when your data starts a few rows down but spans the full width. Combine it with n_max to also cap how many data rows you read.

Using range to read an exact cell block

The range argument names a rectangle of cells in A1 notation, so it controls rows and columns together. Reading range = "A3:C6" starts at row 3 (the header), ends at row 6, and keeps columns A through C. The result matches the clean skip read above.

Because range also picks columns, you can drop unwanted ones. Reading range = "B3:C6" keeps only units and revenue and ignores the product column entirely.

When you want a row range but every column, the cell_rows() helper avoids hard-coding a final column letter. It is the closest range equivalent to a pure skip.

skip vs range: which should you use

Reach for skip when only the row offset matters, and range when you need to box in a region. The table below sums up the trade-offs.

A simple rule: if your only problem is junk rows at the top, use skip. If the sheet holds several tables, notes, or extra columns and you want just one block, use range.

Common pitfalls

Off-by-one errors are the most frequent skip mistake. To reach a header on Excel row 3, you skip the 2 rows above it, so skip = 2. The skip value is the count of rows discarded, not the row number where data begins.

Passing a numeric vector to range fails. The range argument expects an A1 string such as "A3:C6" or a cell_*() helper. A bare range = 3:6 raises an error. Use range = cell_rows(3:6) for a row-only selection.

Skipping rows does not fix column types. Cells stored as text in Excel still read as character vectors. After a clean skip or range read, apply readr::type_convert() or set col_types to get numeric columns.

Try it yourself

Related readxl functions

These helpers extend the row and column control that skip and range give you.

• read_excel(): the main reader that accepts both skip and range.
• cell_rows(): build a range from row numbers only.
• cell_cols(): build a range from column letters only.
• anchored(): define a range by a corner cell plus a width and height.
• excel_sheets(): list sheet names before deciding what to read.

See the official readxl read_excel reference for the full argument list.

FAQ

Does range override skip in read_excel()?

Yes. When you supply range, readxl ignores both skip and n_max. The override is silent, with no error or warning, so passing both arguments is a common source of confusion. If you need a row offset, either use skip alone or express the offset inside the range, for example cell_rows(3:90).

How do I skip rows but keep the header in readxl?

Set skip to the number of rows above the header. With two title rows, skip = 2 discards them, and readxl then treats the next row as the header automatically. You do not need a separate header argument. If there is no header row at all, add col_names = FALSE or pass your own names.

Can range select specific columns only?

Yes, and this is the main reason to prefer range over skip. A range like "B3:D90" keeps only columns B through D. The skip argument cannot drop columns; it always reads the full width of the sheet. Use cell_cols("B:D") if you want those columns for every row.

What is the difference between skip and n_max?

skip removes rows from the top before reading, while n_max caps how many data rows are read after the header. They work together: skip = 2, n_max = 100 ignores two title rows and then reads at most 100 data rows. Both are ignored when range is set.