diff --git a/NEWS.md b/NEWS.md
index 4ddb742..cba74c4 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -4,6 +4,8 @@
     `-F10` (#26 @cablegui).
 * No longer misinterprets date formats that use underscores `_` as dates when
     the underscore is followed by a date-ish character like `M` (#24).
+* Optionally omits blank cells with `include_blank_cells = FALSE` in
+    `xlsx_cells()` (#25).
 
 # tidyxl 1.0.1
 
diff --git a/R/RcppExports.R b/R/RcppExports.R
index 3eebe27..fbef21c 100644
--- a/R/RcppExports.R
+++ b/R/RcppExports.R
@@ -1,8 +1,8 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
-xlsx_cells_ <- function(path, sheet_paths, sheet_names, comments_paths) {
-    .Call('_tidyxl_xlsx_cells_', PACKAGE = 'tidyxl', path, sheet_paths, sheet_names, comments_paths)
+xlsx_cells_ <- function(path, sheet_paths, sheet_names, comments_paths, include_blank_cells) {
+    .Call('_tidyxl_xlsx_cells_', PACKAGE = 'tidyxl', path, sheet_paths, sheet_names, comments_paths, include_blank_cells)
 }
 
 xlsx_formats_ <- function(path) {
diff --git a/R/tidy_xlsx.R b/R/tidy_xlsx.R
index 651dc80..43b5454 100644
--- a/R/tidy_xlsx.R
+++ b/R/tidy_xlsx.R
@@ -193,7 +193,7 @@ tidy_xlsx <- function(path, sheets = NA) {
   all_sheets <- utils_xlsx_sheet_files(path)
   sheets <- check_sheets(sheets, path)
   formats <- xlsx_formats_(path)
-  cells <- xlsx_cells_(path, sheets$sheet_path, sheets$name, sheets$comments_path)
+  cells <- xlsx_cells_(path, sheets$sheet_path, sheets$name, sheets$comments_path, include_blank_cells = TRUE)
   # Split into a list of data frames, one per sheet
   cells$sheet <- factor(cells$sheet, levels = sheets$name) # control sheet order
   cells_list <- split(cells, cells$sheet)
diff --git a/R/xlsx_cells.R b/R/xlsx_cells.R
index f4ffca3..c0ec014 100644
--- a/R/xlsx_cells.R
+++ b/R/xlsx_cells.R
@@ -13,6 +13,10 @@
 #' @param check_filetype Logical. Whether to check that the filetype is xlsx (or
 #' xlsm) by looking at the file itself, rather than using the filename
 #' extension.
+#' @param include_blank_cells Logical. Whether to include cells that have no
+#' value or formula (but might have formatting or comments).  Useful when a
+#' whole column of cells has been formatted, but most are empty.  Try setting
+#' this to `FALSE` if a spreadsheet seems too large to load.
 #'
 #' @return
 #' A data frame with the following columns.
@@ -138,11 +142,13 @@
 #' # In-cell formatting is available in the `character_formatted` column as a
 #' # data frame, one row per substring.
 #' xlsx_cells(examples)$character_formatted[77]
-xlsx_cells <- function(path, sheets = NA, check_filetype = TRUE) {
+xlsx_cells <- function(path, sheets = NA, check_filetype = TRUE,
+                       include_blank_cells = TRUE) {
   path <- check_file(path)
   sheets <- check_sheets(sheets, path)
   xlsx_cells_(path,
               sheets$sheet_path,
               sheets$name,
-              sheets$comments_path)
+              sheets$comments_path,
+              include_blank_cells)
 }
diff --git a/man/xlsx_cells.Rd b/man/xlsx_cells.Rd
index 1aec41d..3336795 100644
--- a/man/xlsx_cells.Rd
+++ b/man/xlsx_cells.Rd
@@ -4,7 +4,8 @@
 \alias{xlsx_cells}
 \title{Import xlsx (Excel) cell contents into a tidy structure.}
 \usage{
-xlsx_cells(path, sheets = NA, check_filetype = TRUE)
+xlsx_cells(path, sheets = NA, check_filetype = TRUE,
+  include_blank_cells = TRUE)
 }
 \arguments{
 \item{path}{Path to the xlsx file.}
@@ -16,6 +17,11 @@ sheets).}
 \item{check_filetype}{Logical. Whether to check that the filetype is xlsx (or
 xlsm) by looking at the file itself, rather than using the filename
 extension.}
+
+\item{include_blank_cells}{Logical. Whether to include cells that have no
+value or formula (but might have formatting or comments).  Useful when a
+whole column of cells has been formatted, but most are empty.  Try setting
+this to \code{FALSE} if a spreadsheet seems too large to load.}
 }
 \value{
 A data frame with the following columns.
diff --git a/src/RcppExports.cpp b/src/RcppExports.cpp
index 62f1fbe..3000fd0 100644
--- a/src/RcppExports.cpp
+++ b/src/RcppExports.cpp
@@ -6,8 +6,8 @@
 using namespace Rcpp;
 
 // xlsx_cells_
-List xlsx_cells_(std::string path, CharacterVector sheet_paths, CharacterVector sheet_names, CharacterVector comments_paths);
-RcppExport SEXP _tidyxl_xlsx_cells_(SEXP pathSEXP, SEXP sheet_pathsSEXP, SEXP sheet_namesSEXP, SEXP comments_pathsSEXP) {
+List xlsx_cells_(std::string path, CharacterVector sheet_paths, CharacterVector sheet_names, CharacterVector comments_paths, bool include_blank_cells);
+RcppExport SEXP _tidyxl_xlsx_cells_(SEXP pathSEXP, SEXP sheet_pathsSEXP, SEXP sheet_namesSEXP, SEXP comments_pathsSEXP, SEXP include_blank_cellsSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
@@ -15,7 +15,8 @@ BEGIN_RCPP
     Rcpp::traits::input_parameter< CharacterVector >::type sheet_paths(sheet_pathsSEXP);
     Rcpp::traits::input_parameter< CharacterVector >::type sheet_names(sheet_namesSEXP);
     Rcpp::traits::input_parameter< CharacterVector >::type comments_paths(comments_pathsSEXP);
-    rcpp_result_gen = Rcpp::wrap(xlsx_cells_(path, sheet_paths, sheet_names, comments_paths));
+    Rcpp::traits::input_parameter< bool >::type include_blank_cells(include_blank_cellsSEXP);
+    rcpp_result_gen = Rcpp::wrap(xlsx_cells_(path, sheet_paths, sheet_names, comments_paths, include_blank_cells));
     return rcpp_result_gen;
 END_RCPP
 }
@@ -100,7 +101,7 @@ END_RCPP
 }
 
 static const R_CallMethodDef CallEntries[] = {
-    {"_tidyxl_xlsx_cells_", (DL_FUNC) &_tidyxl_xlsx_cells_, 4},
+    {"_tidyxl_xlsx_cells_", (DL_FUNC) &_tidyxl_xlsx_cells_, 5},
     {"_tidyxl_xlsx_formats_", (DL_FUNC) &_tidyxl_xlsx_formats_, 1},
     {"_tidyxl_xlsx_sheet_files_", (DL_FUNC) &_tidyxl_xlsx_sheet_files_, 1},
     {"_tidyxl_xlsx_validation_", (DL_FUNC) &_tidyxl_xlsx_validation_, 3},
diff --git a/src/tidyxl.cpp b/src/tidyxl.cpp
index 5d19352..c95c974 100644
--- a/src/tidyxl.cpp
+++ b/src/tidyxl.cpp
@@ -21,9 +21,11 @@ List xlsx_cells_(
     std::string path,
     CharacterVector sheet_paths,
     CharacterVector sheet_names,
-    CharacterVector comments_paths
+    CharacterVector comments_paths,
+    bool include_blank_cells
     ) {
-  xlsxbook book(path, sheet_paths, sheet_names, comments_paths);
+  xlsxbook book(path, sheet_paths, sheet_names, comments_paths,
+      include_blank_cells);
   return book.information_;
 }
 
diff --git a/src/xlsxbook.cpp b/src/xlsxbook.cpp
index 8832bee..31134f9 100644
--- a/src/xlsxbook.cpp
+++ b/src/xlsxbook.cpp
@@ -24,7 +24,8 @@ xlsxbook::xlsxbook(
     const std::string& path,
     CharacterVector& sheet_paths,
     CharacterVector& sheet_names,
-    CharacterVector& comments_paths):
+    CharacterVector& comments_paths,
+    const bool& include_blank_cells):
   path_(path),
   sheet_paths_(sheet_paths),
   sheet_names_(sheet_names),
@@ -43,7 +44,7 @@ xlsxbook::xlsxbook(
   createSheets();
   countCells();
   initializeColumns();
-  cacheInformation();
+  cacheInformation(include_blank_cells);
 }
 
 // Based on hadley/readxl
@@ -157,7 +158,7 @@ void xlsxbook::initializeColumns() {
   local_format_id_ = IntegerVector(cellcount_,   NA_INTEGER);
 }
 
-void xlsxbook::cacheInformation() {
+void xlsxbook::cacheInformation(const bool& include_blank_cells) {
   // Loop through sheets
   List sheet_list(sheet_paths_.size());
 
@@ -175,7 +176,7 @@ void xlsxbook::cacheInformation() {
     doc.parse<rapidxml::parse_strip_xml_namespaces>(&(*xml)[0]);
     rapidxml::xml_node<>* workbook = doc.first_node("worksheet");
     rapidxml::xml_node<>* sheetData = workbook->first_node("sheetData");
-    sheet->parseSheetData(sheetData, i);
+    sheet->parseSheetData(sheetData, i, include_blank_cells);
     sheet->appendComments(i);
   }
 
diff --git a/src/xlsxbook.h b/src/xlsxbook.h
index f89ef00..224e4c7 100644
--- a/src/xlsxbook.h
+++ b/src/xlsxbook.h
@@ -57,7 +57,8 @@ class xlsxbook {
         const std::string& path,
         Rcpp::CharacterVector& sheet_names,
         Rcpp::CharacterVector& sheet_paths,
-        Rcpp::CharacterVector& comments_paths
+        Rcpp::CharacterVector& comments_paths,
+        const bool& include_blank_cells
         );
 
     void cacheStrings();
@@ -67,7 +68,7 @@ class xlsxbook {
     void countCells();
     void initializeColumns();
     void cacheCells();
-    void cacheInformation();
+    void cacheInformation(const bool& include_blank_cells);
 
 };
 
diff --git a/src/xlsxcell.cpp b/src/xlsxcell.cpp
index 96ee002..c287334 100644
--- a/src/xlsxcell.cpp
+++ b/src/xlsxcell.cpp
@@ -68,6 +68,7 @@ void xlsxcell::cacheValue(
   if (v != NULL) {
     vvalue = v->value();
   } else {
+    // TODO: don't construct the cell
     book.is_blank_[i] = true;
   }
 
diff --git a/src/xlsxsheet.cpp b/src/xlsxsheet.cpp
index a25879e..e7c70b1 100644
--- a/src/xlsxsheet.cpp
+++ b/src/xlsxsheet.cpp
@@ -141,7 +141,8 @@ void xlsxsheet::cacheComments(String comments_path) {
 
 void xlsxsheet::parseSheetData(
     rapidxml::xml_node<>* sheetData,
-    unsigned long long int& i) {
+    unsigned long long int& i,
+    const bool& include_blank_cells) {
   // Iterate through rows and cells in sheetData.  Cell elements are children
   // of row elements.  Columns are described elswhere in cols->col.
   rowHeights_.assign(1048576, defaultRowHeight_); // cache rowHeight while here
@@ -160,19 +161,43 @@ void xlsxsheet::parseSheetData(
       rowHeights_[rowNumber - 1] = rowHeight;
     }
 
-    for (rapidxml::xml_node<>* c = row->first_node();
-        c; c = c->next_sibling()) {
-      xlsxcell cell(c, this, book_, i);
+    if (include_blank_cells) {
+      for (rapidxml::xml_node<>* c = row->first_node();
+          c; c = c->next_sibling()) {
+        xlsxcell cell(c, this, book_, i);
 
-      // Sheet name, row height and col width aren't really determined by the
-      // cell, so they're done in this sheet instance
-      book_.sheet_[i] = name_;
-      book_.height_[i] = rowHeight;
-      book_.width_[i] = colWidths_[book_.col_[i] - 1];
+        // Sheet name, row height and col width aren't really determined by
+        // the cell, so they're done in this sheet instance
+        book_.sheet_[i] = name_;
+        book_.height_[i] = rowHeight;
+        book_.width_[i] = colWidths_[book_.col_[i] - 1];
 
-      ++i;
-      if ((i + 1) % 1000 == 0)
-        checkUserInterrupt();
+        ++i;
+        if ((i + 1) % 1000 == 0)
+          checkUserInterrupt();
+      }
+    } else {
+      for (rapidxml::xml_node<>* c = row->first_node();
+          c; c = c->next_sibling()) {
+        // If cell has no child nodes then it is empty (no value or formula)
+        // besides maybe formatting (linked to via attributes not child nodes).
+        rapidxml::xml_node<>* first_child = c->first_node();
+        if (first_child != NULL) {
+          xlsxcell cell(c, this, book_, i);
+
+          // TODO: check readxl's method of importing ranges
+
+          // Sheet name, row height and col width aren't really determined by
+          // the cell, so they're done in this sheet instance
+          book_.sheet_[i] = name_;
+          book_.height_[i] = rowHeight;
+          book_.width_[i] = colWidths_[book_.col_[i] - 1];
+
+          ++i;
+          if ((i + 1) % 1000 == 0)
+            checkUserInterrupt();
+        }
+      }
     }
   }
 }
diff --git a/src/xlsxsheet.h b/src/xlsxsheet.h
index 867ee6b..8cad0fc 100644
--- a/src/xlsxsheet.h
+++ b/src/xlsxsheet.h
@@ -35,7 +35,8 @@ class xlsxsheet {
     void cacheComments(Rcpp::String comments_path);
     void parseSheetData(
         rapidxml::xml_node<>* sheetData,
-        unsigned long long int& i);
+        unsigned long long int& i,
+        const bool& include_blank_cells);
     void appendComments(unsigned long long int& i);
 
 };
diff --git a/tests/testthat/test-xlsx_cells.R b/tests/testthat/test-xlsx_cells.R
index 44bd337..26c9649 100644
--- a/tests/testthat/test-xlsx_cells.R
+++ b/tests/testthat/test-xlsx_cells.R
@@ -39,3 +39,11 @@ test_that("array formulas are detected as such", {
   expect_equal(cells$is_array[43], TRUE)
   expect_equal(cells$is_array[45], TRUE)
 })
+
+test_that("include_blank_cells works", {
+  cells <- xlsx_cells("./examples.xlsx", include_blank_cells = FALSE)
+  blanks <- cells[cells$is_blank, ]
+  non_blanks <- cells[!cells$is_blank, ]
+  expect_equal(nrow(blanks), 0L)
+  expect_gt(nrow(non_blanks), 0L)
+})