忽略单元格的顺序，删除Google表格中的重复行

Question

I have a spreadsheet with many "duplicate" rows. I don't want to remove them manually, since I have thousands of rows. In my particular case, the rows may not be exact duplicates , as I don't care about the order of cells in the rows. Here is an example:

A   B
dog cat
apple orange
red blue
cat dog

dog cat and cat dog are duplicate in my case. So, I want to keep one of them. Don't care which one, could be the first or last.

I know I need some kind of order-independent row comparison. How can this be accomplished using spreadsheet formulas or Google Apps Script?

PS My actual data has 7 columns and not 2 as my example. (A to G)

Answer 1

I don't have experience with Google Sheets, but heres what I would do in Excel and hopefully you can replicate this in some way.

As a comment suggests, using Remove Duplicates will do as you require, but first you need to normalise the list in a way that will pick up these duplicates.

In Column C: =IF(A1<B1,A1,B1)

In Column D: =IF(A1<B1,B1,A1)

This will essentially put the values in Columns A & B in alphabetical order between columns C & D, you can then perform the Remove Duplicates on these two new columns.

Answer 2

根据奥利弗·卡尔（Oliver Carr）提供的解决方案，这里是单配方解决方案：

=unique(arrayformula({IF(A:A<B:B,A:A,B:B),IF(A:A<B:B,B:B,A:A)}))

Answer 3

For a flexible solution for many rows and maybe more columns, try this:

Assuming your data is in columns A:G, starting in row 1.

In H1: =2^COUNTIF($A:$G,"<"&A1) Copy this formula over from H to N, and down as many rows as needed to cover all cells of data.

In O1: =SUM($H:$N) and copy this down to cover all rows.

The value in O will be the same for rows with the same words. You can now do remove duplicates on column O.

What you are doing is assigning every word in your data a unique power of 2, and the sum is guaranteed to be unique (think of a binary number with as many digits as you have unique words, with each digit set to 1 if that word appears in the row, and otherwise zero).

If you want a more condensed version, this can also be entered as an array formula (again with data in A:G):

In H1: =SUM(2^COUNTIF(A:G,"<"&A1:G1)) enter by pressing Cntl+Shift+Enter Or in Google Sheets: =ArrayFormula(SUM(2^COUNTIF(A:C,"<"&A1:C1)))

If you are restricted to Google Sheets, you can then use the UNIQUE function on H to get a unique list of IDs, and perform lookups into the original table to get the actual rows.

Answer 4

The answers by Spencer , Max and Oliver all use spreadsheet formulas to return filtered arrays. They have the advantage that they can automatically recalculate when new rows are added to the source data.

However, you asked specifically about deleting rows, which none of those answers do. To accomplish this, you must use a script, as formulas do not remove source data from spreadsheets.

This snippet contains a complete script, including a menu-driven user interface that will invoke the delSimilarRows() function for the current sheet. It is written to be a spreadsheet-contained script, but could be easily converted to be an add-on.

 /** * @OnlyCurrentDoc Limits the script to only accessing the current spreadsheet. */ /** * Adds a custom menu * * @param {Object} e The event parameter for a simple onOpen trigger. */ function onOpen(e) { SpreadsheetApp.getUi() .createMenu('Custom') .addItem('Delete similar rows', 'delSimRowsGUI') .addToUi(); } /** * Prompt user for confirmation before proceeding with deletion. * Provide results after operation. * */ function delSimRowsGUI() { var ui = SpreadsheetApp.getUi(); var choice = ui.alert("Confirm action", "This will delete rows in the current sheet that contain sets of cells that already appear together in other rows.", ui.ButtonSet.OK_CANCEL); if (choice === ui.Button.OK) { var numDeleted = delSimilarRows(); ui.alert("Deleted "+numDeleted+" row"+(numDeleted===1?'.':'s.')); } } /** * Delete rows in the current sheet that contain sets of cells that already * appear together in other rows. (Almost duplicates, but order-independent.) * From: https://stackoverflow.com/a/37304191/1677912 * * @returns {Number} The number of matching rows that were deleted. */ function delSimilarRows() { // Get all rows from sheet. var currentSheet = SpreadsheetApp.getActiveSheet(); var data = currentSheet.getDataRange().getValues(); var numDeleted = 0; // Sort cells within rows, and join into a string with (hopefully!) unique delimiter var sorted = data.map(function(row) { return row.sort().join(' |-| '); }); // Identify duplicate rows in the sorted data, and delete the corresponding // spreadsheet rows. (Note: looping backwards, so deletion is clean.) for (var row=sorted.length-1; row>=0; row--) { if (sorted.slice(0,row).indexOf(sorted[row]) !== -1) { currentSheet.deleteRow(row+1); numDeleted++; } } return numDeleted; } 

The function that does all the real work is delSimilarRows() . It uses some JavaScript magic to identify rows to be removed, and directly deletes them from the current sheet.

It handles all types of data, by temporarily converting rows to their string representations, with cell contents sorted alphabetically and a (hopefully) unique separator between them. Doing this, your example data will appear (to the computer only) like this:

[ "cat |-| dog",
  "apple |-| orange",
  "blue |-| red",
  "cat |-| dog" ]

We can then loop through the rows checking for duplicates using the JavaScript Array.indexOf() method on slices of the row array excluding our current row.

Because we're dealing with 0-based JavaScript arrays as well as 1-based Spreadsheet rows, we need to be careful about adding or subtracting 1 while indexing one or the other.

/**
 * Delete rows in the current sheet that contain sets of cells that already 
 * appear together in other rows. (Almost duplicates, but order-independent.)
 * From: https://stackoverflow.com/a/37304191/1677912
 *
 * @returns {Number}       The number of matching rows that were deleted.
 */
function delSimilarRows() {
  // Get all rows from sheet.
  var currentSheet = SpreadsheetApp.getActiveSheet();
  var data = currentSheet.getDataRange().getValues();
  var numDeleted = 0;

  // Sort cells within rows, and join into a string with (hopefully!) unique delimiter
  var sorted = data.map(function(row) {
    return row.sort().join(' |-| ');
  });

  // Identify duplicate rows in the sorted data, and delete the corresponding
  // spreadsheet rows. (Note: looping backwards, so deletion is clean.)
  for (var row=sorted.length-1; row>=0; row--) {
    if (sorted.slice(0,row).indexOf(sorted[row]) !== -1) {
      currentSheet.deleteRow(row+1);
      numDeleted++;
    }
  }
  return numDeleted;
}