ViewCSV extension help

A file won't open

• Larger than 52,428,800 bytes (50 MB): split the file or use a desktop tool.
• "This is an Excel or other spreadsheet workbook": save it as CSV UTF-8 in your spreadsheet app and open that file.
• "This looks like a binary file": compressed (.gz) files must be unpacked first; PDF isn't supported.
• More than 1,000,000 rows, 1,000 columns or 5,000,000 cells: if the column count looks wrong, the delimiter is probably wrong (see below).

Letters look wrong or the file "isn't valid UTF-8"

ViewCSV detects UTF-8 and UTF-16 with a byte-order mark. Older spreadsheet exports often use another encoding: Excel's plain "CSV" on Western European Windows usually saves Windows-1252, while "CSV UTF-8" saves UTF-8.

1. Open Import settings (on the start page, or in the top bar once a file is open).
2. Choose the encoding the file was saved in, for example Windows-1252, Windows-1250, Windows-1251, Shift_JIS, EUC-KR or GB18030.
3. Open the file again, or choose Read again. Nothing is converted. Bytes that don't fit UTF-8 or UTF-16 give an error; other encodings can accept the wrong bytes without an error, so if letters still look wrong, try another encoding.

Everything is in one column

The delimiter was not recognized. In Import settings, choose comma, semicolon, tab or pipe, then read the file again. Files saved with a semicolon and decimal comma (common in many European spreadsheets) open fine; the numbers are simply kept as text, such as 19,90.

"A quoted value … never ends"

A double quote is missing in the file, starting at the line shown in the message. Fix the file if you can. To see it anyway, turn off Treat double quotes as CSV quoting in Import settings; quotes are then read as ordinary characters. For pasted text, the message offers Open with quotes as text directly.

Pasting

• The toolbar popup accepts up to 4,194,304 characters. For longer text, open the ViewCSV tab and choose Paste text there (up to 52,428,800 characters).
• Rows copied from a spreadsheet are usually tab-separated; ViewCSV detects that.
• Empty lines are skipped and counted in a notice above the table.

Find tables shows nothing, or not the table you see

• Only real HTML tables are found. Many sites draw grids with other elements; those can't be read. Copy the text and paste it instead, or look for a CSV download on the site.
• Tables inside frames from another site, hidden tables and one-cell layout tables are skipped. At most 50 tables are listed.
• If the page loads more rows as you scroll or page through, only the rows on the page at that moment are read.
• "The table on the page has changed" or "can no longer read that page": the page was reloaded or changed. Go back to it and choose Find tables again.
• Merged cells are split and their value is repeated in each cell they covered; a notice tells you how many.

"ViewCSV can't read this page"

Browsers don't let extensions read their own pages (settings, new tab, extension store) or some protected pages. Files on your computer opened in a tab are also off limits: open them with Choose file in the ViewCSV tab.

A CSV link doesn't open

• Access wasn't allowed: choose Allow and open again and accept the browser's question for that site.
• "The link redirects to another address": ViewCSV doesn't follow redirects. Open the link in your browser, then use the final address, or download the file.
• "Requires you to sign in" (error 401 or 403): ViewCSV doesn't send your sign-in. Download the file in your browser instead.
• "Returned a web page, not a CSV file": the link leads to a sign-in or download page rather than the file.
• Too large or slower than 60 seconds, a user name or password in the link, or an address on your own network: the download is refused.
• In every case you can download the file with your browser and open it with Choose file.

Exports

• Exports are UTF-8. Turn on the UTF-8 BOM option (on by default) if a spreadsheet shows accents or non-Latin text incorrectly.
• CSV export uses the file's own delimiter by default; choose another in the Export dialog. Values are written exactly as they are, not reformatted.
• Formula protection (on by default) adds an apostrophe before cells that start with =, +, - or @ (or their full-width forms, a tab or a carriage return) and aren't plain numbers, so a spreadsheet shows them as text instead of running them. Turn it off if you need the exact original text, for example to load the file into a database.
• JSON export is an array of objects with every value as a string. Unnamed columns become "Column N" and repeated names get "(2)", "(3)".
• Choose whether to export the rows shown (after search and filters, in the current order) or all rows, and whether to leave out hidden columns.
• Exports larger than 209,715,200 bytes (200 MB) are refused: filter rows or hide columns first.

The original file and unsaved changes

ViewCSV never changes your original file or the web page. Edits, deleted rows and renamed columns exist only in the ViewCSV tab until you export, and Undo can reverse the last 200 changes. Sorting, searching and filtering only change what you see. Closing or reloading the tab discards changes (the browser asks first), and changing Import settings reads the data again and discards them too.

A number column sorts like text

Numbers are recognized only as plain decimals with a dot, such as 1234.5 or -0.75. Decimal commas, thousands separators, currency symbols, exponents and dates are text, and a column with leading zeros (postal codes, IDs) is text on purpose. Text is sorted in natural order, so 2 still comes before 10.