如何对数据条目进行排序，以便在 Python 或 Google 表格中将一系列数据组合在一起？

Question

For instance, given a dataset of various weights and names, how can I group individuals with similar weights (+/- 5% the weight I want) together?

Thank you!

Answer 1

I think you can use package. For excel, openpyxl is such essential package in python. You can do sort, data entry, making chart on excel by using this package in python. First, go to this link.

https://pypi.org/ search openpyxl and install this package.

https://openpyxl.readthedocs.io/en/stable/tutorial.html#create-a-workbook this link is for guideline to use openpyxl.

Answer 2

I understand that you want to divide your data in groups where the maximum difference from the biggest to the smaller is ±5%. I wrote an Apps Script code that can do that in your Sheet. First, I wrote an example Sheet with student names and test scores (from 0 to 10) instead of names and heights; I'll explain why later. This is the initial state of the example sheet:

In the Group column the code will drop the group ID as an integer starting at 0. This is the code:

function so62060595() {
  var dataColumn = 2; // Column B
  var groupColumn = 3; // Column C
  var dataSheet = SpreadsheetApp.getActive().getActiveSheet();
  var dataRange = dataSheet.getRange(2, 1, dataSheet.getLastRow() - 1, dataSheet
    .getLastColumn()).sort({
    column: dataColumn,
    ascending: false
  });
  var data = dataRange.getValues();
  var groupingPercentage = 5 / 100 // 5%
  var upperBound = data[0][dataColumn - 1];
  var groupID = 0;

  for (var r = 0; r < data.length; r++) {
    if (upperBound - upperBound * groupingPercentage <= data[r][dataColumn -
      1] + data[r][dataColumn - 1] * groupingPercentage) {
      // Include in the same group
      data[r][groupColumn - 1] = groupID;
    } else {
      // Create a new group
      var groupID = groupID + 1;
      var upperBound = data[r][dataColumn - 1];

      data[r][groupColumn - 1] = groupID;
    }
  }

  dataRange.setValues(data);
}

The first thing that the code does is to open the sheet with SpreadsheetApp.getActive() , Spreadsheet.getActiveSheet() . After that, the code looks for the data range with Sheet.getRange() (notice how it uses Sheet.getLastRow() and Sheet.getLastColumn() to find the range size) and sorts it with Range.sort() using the score column as a reference. Later, it reads the range with Range.getValues() . Also, I initialized some variables like the ID and group columns, the desired grouping percentage (5%) in this case and the initial group ID ( 0 ).

After all that initialization the code will iterate over every row and check if the data value ( Score in the example) is at a ±5% of distance to the upper bound of the group (the highest value in the group). If the value is in the ±5% range, the group ID will be dropped. If it's not in the range, a new group ID will be generated and the upper bound taken from that entry. The process will continue until all entries have a group ID, and after that the data will be entered in the table with Range.setValues() . The final result looks like this:

And now, why did I use test scores instead of heights? Well, look what happens with an example of heights using the previous code:

Only two groups are generated ( 0 and 1 ) because the distance between realistic heights is smaller than ±5%. I hope that my answer helps you, but don't hesitate to ask me any additional doubt.

---

UPDATE

Based on the question update in your comment I modified the script. If I understand correctly, you want a midpoint in each group and calculate the group bounds based on that midpoint plus/minus 5%. If my assumption is correct, you can use the following code:

function calculateGroupBounds(groupingPercentage, groupUpperBound) {
  var groupBounds = {};
  groupBounds['groupUpperBound'] = groupUpperBound;
  groupBounds['groupMidpoint'] = 100 * groupUpperBound / (100 +
    groupingPercentage);
  groupBounds['groupLowerBound'] = (100 - groupingPercentage) * groupBounds[
    'groupMidpoint'] / 100;

  return groupBounds;
}

function so62060595B() {
  // Sheet reading
  var dataColumn = 2; // Column B
  var groupColumn = 3; // Column C
  var dataSheet = SpreadsheetApp.getActive().getActiveSheet();
  var dataRange = dataSheet.getRange(2, 1, dataSheet.getLastRow() - 1, dataSheet
    .getLastColumn()).sort({
    column: dataColumn,
    ascending: false
  });
  var data = dataRange.getValues();

  // Group initialization
  var groupingPercentage = 5; // 5%
  var groupID = 0;
  var groupUpperBound = data[0][dataColumn - 1];
  var groupBounds = calculateGroupBounds(groupingPercentage, groupUpperBound)
  var groupMidpoint = groupBounds['groupMidpoint'];
  var groupLowerBound = groupBounds['groupLowerBound'];

  for (var r = 0; r < data.length; r++) {
    if (data[r][dataColumn - 1] <= groupUpperBound && data[r][dataColumn - 1] >=
      groupLowerBound) {
      // Include in the same group
      data[r][groupColumn - 1] = groupID;
    } else {
      // Create a new group
      var groupID = groupID + 1;
      var upperBound = data[r][dataColumn - 1];
      var groupBounds = calculateGroupBounds(groupingPercentage,
        upperBound)
      var groupMidpoint = groupBounds['groupMidpoint'];
      var groupLowerBound = groupBounds['groupLowerBound'];

      data[r][groupColumn - 1] = groupID;
    }
  }

  dataRange.setValues(data);
}

This new code uses the same Apps Script methods as the previous one and includes a new function ( calculateGroupBounds() ) to compute the upper and lower bounds and the midpoint. In the data iteration the code will check if the value falls between the upper and lower bound, and if it does the group iD will be dropped. If it doesn't a new group will be created. This is the result with the same example data as the previous code:

And this is the result with the heights table:

These results are the same as the previous code even though we use a different approach in this second code. This is because on the first code I used mathematical properties to divide the data in ±5% groups without calculating the midpoint. Please, ask me any doubts if you still need help.