{"id":45923,"date":"2026-11-12T05:25:59","date_gmt":"2026-11-12T05:25:59","guid":{"rendered":"https:\/\/www.withdata.com\/blog\/?p=45923"},"modified":"2026-11-12T05:43:12","modified_gmt":"2026-11-12T05:43:12","slug":"excel-wont-open-large-file-row-limit-and-what-opens-it","status":"publish","type":"post","link":"https:\/\/www.withdata.com\/blog\/bigdatafileviewer\/excel-wont-open-large-file-row-limit-and-what-opens-it.html","title":{"rendered":"Excel Won’t Open a 350 MB File: The Row Ceiling, the Memory Wall, and What Actually Opens It"},"content":{"rendered":"

The workbook is on your disk. Double-click it, and Excel either shows “Excel cannot complete this task with available resources”<\/em>, sits at 0% until the file stops responding, or opens 1,048,576 rows and quietly stops. The file is not corrupt, and re-saving it from the machine that made it will not help:<\/p>\n\n\n\n\n\n\n\n\n
Limit<\/th>\nValue<\/th>\nWhat it means<\/th>\n<\/tr>\n<\/thead>\n
Rows per worksheet<\/td>\n1,048,576<\/strong><\/td>\na .xlsx<\/code> cannot hold more, whatever your hardware<\/td>\n<\/tr>\n
Columns per worksheet<\/td>\n16,384<\/td>\nrarely the wall<\/td>\n<\/tr>\n
Workbook memory<\/td>\nroughly 10\u00d7<\/strong> the file size<\/td>\n400 MB of xlsx easily becomes 4 GB of RAM in Excel<\/td>\n<\/tr>\n
32-bit Office<\/td>\n2 GB address space<\/td>\nit cannot get near that, so it dies first<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The people in the threads say the same thing. In r\/excel, “Excel file far too large to even open”<\/em> (2026-04, 51 comments) and “Excel grew from 400MB to 650MB”<\/em> (2026-06, 77 comments) are both about the same wall, and in “Having trouble extrapolating data from a 10GB CSV file”<\/em> the advice is to stop using Excel for the job at all.<\/p>\n

1. Three routes out, and only one shows you the data<\/h2>\n

Route one \u2014 make the workbook smaller.<\/strong> Delete unused columns, drop formatting, split it into twelve workbooks and puzzle it back together in your head. This is the route everyone starts with, and it is work you repeat every time the report is regenerated.<\/p>\n

Route two \u2014 convert it and query it.<\/strong> Convert the xlsx to CSV and open that in a database, Power Query or pandas<\/code>. It works, and if you already live in SQL it is a good answer \u2014 but you no longer have the thing you wanted: the file, on screen, scrolled to row 900,000.<\/p>\n

Route three \u2014 open it in something that is not Excel.<\/strong> A grid viewer reads the workbook as a stream of rows instead of building the whole object model in memory, so the file opens and the memory stays flat. That is what BigDataFileViewer<\/a> does; it is a read-only viewer, which is the point \u2014 you are looking at a 400 MB report, not editing it.<\/p>\n

2. Opening it \u2014 the GUI route<\/h2>\n
    \n
  1. File \u2192 Open File…<\/strong> (Ctrl+O), or drag the .xlsx<\/code> onto the window. The file opens in its own tab.<\/li>\n
  2. The Sheet<\/strong> drop-down at the top picks the worksheet; the status bar shows the shape:<\/li>\n<\/ol>\n

    rows \u00d7 cols<\/code>, the file size, how long it took to open, and the memory it is using.<\/p>\n

      \n
    1. Go to row<\/strong> (Ctrl+G) takes a row number; Ctrl+Home and Ctrl+End go to the first and last row.<\/li>\n<\/ol>\n

      The grid keeps one page of 50,000 rows in memory and moves the page as you scroll, so the row count does not change the memory figure.<\/p>\n

        \n
      1. Search \u2192 Find…<\/strong> (Ctrl+F) scans the whole workbook and lists the hits; double-click one to jump<\/li>\n<\/ol>\n

        to it. For .xlsx<\/code> this means decompressing the sheet, so it is slower than for CSV, and the progress line says how far it has got.<\/p>\n

          \n
        1. File \u2192 Export View as CSV…<\/strong> writes what you are looking at as CSV or TSV, header row included \u2014<\/li>\n<\/ol>\n

          which is how you hand a slice of the 400 MB file to someone who only has Excel.<\/p>\n

          \"BigDataFileViewer<\/p>\n

          Measured on the machine that wrote this (Windows, 64-bit), with files opened through the GUI above:<\/p>\n\n\n\n\n\n\n\n
          File<\/th>\nRows \u00d7 columns<\/th>\nOpen<\/th>\nMemory<\/th>\nJump to a random row<\/th>\n<\/tr>\n<\/thead>\n
          .xlsx<\/code>, 345 MB<\/td>\n1,048,575 \u00d7 14<\/td>\n9.9 s<\/strong><\/td>\n119 MB<\/td>\nlast page, still < 1 s<\/td>\n<\/tr>\n
          .xlsx<\/code>, 236 MB<\/td>\n800,000 \u00d7 8<\/td>\n0.72 s<\/strong><\/td>\n21.6 MB<\/td>\nworst of 12 jumps: 0.69 s<\/td>\n<\/tr>\n
          .csv<\/code>, 817 MB<\/td>\n12,000,000 \u00d7 8<\/td>\n1.7 s<\/td>\n21 MB<\/td>\n15 ms average<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

          Excel refuses the first one outright \u2014 1,048,575 data rows plus the header is exactly the row ceiling, and at that size it cannot keep the workbook in memory anyway.<\/p>\n

          3. What you cannot expect from a viewer<\/h2>\n