Numbers: Preserve Row , Preserve Column- do not preserve values in cells ?

If you mean you want your formula to continue using a previous value from a cell after you have changed that value, that is not how spreadsheets work. There is no memory of previous values. If you need to keep track of previous values, they have to be manually saved in a column or row of a table and the formula needs to refer to them there. Typically this other column or row is where you would enter the values in the first place, rather than entering them in some other cell then copy/pasting each day (or whenever the value needs to change).

If you mean you want to take a cell that has a formula in it and convert it into a value that will not change, you can use Copy followed by Paste Formula Results to remove the formula and replace it with its calculated value.

> Note that in this case "price" and "factor" refer to the entire column (see how the both columns are colored). For the given formula, Numbers will use only the cells in the same row as the formula (=A2*B2) but for something like =MAX(price) it will use the entire column.

I kinda have a problem with this. I get the benefit - makes it super easy to deal with row-based formulas like price X factor, but I have a hard time getting past the fact that price and factor are column references, so it should (logically?) refer to the column, not just one row in the column.

As much as it hurts to say, I prefer Excel's model of @[column] which more clearly delineates the intersection vs, entire column reference.

The basic premise is if the parameter can be a range then it will be a range, if it can only be a single cell then it will be a single cell from the same row as the formula. It would be more clear if it highlighted just the cell, not the entire column, when it was going to be referring to a single cell. Internally it knows when "B" means "B2" but it doesn't show that to us.

> The basic premise is if the parameter can be a range then it will be a range, if it can only be a single cell then it will be a single cell from the same row as the formula.

I understand how it works. I was mostly bemoaning the fact that the user interface is inconsistent - 'B' could either mean the entire column B, or the current row in column B, and you can't tell by looking at the function quite what it will do, and that's a flaw.

Case in point: MAX() can take both individual values and ranges - does MAX(B,C) mean the greatest value in columns B and C, or does it mean the greatest value in the current row of columns B and C? (hint: it's the former, but you likely had to think about that)

In contrast, Excel uses @B2:B20 to indicate the corresponding row in column B, which makes sense, and wouldn't be hard to implement (indeed, Numbers exports @column notation if you export an intersection (in Excel's terms) to Excel).

If you want a cell that shows the formula of a cell, use FORMULATEXT()

This function will return a text representation of the formula in the referenced cell.

However, it's still dynamic and changes to the referenced cell will update this value.

Seems to me that cell addresses are less risky than "variable names". With cell addresses you know exactly what cell is being refered to.

Numbers will create names for cells if you have "use header names as labels" turned on (which is the default). If a table has a header row and a header column, each cell will get a name that is the values in its header row and column. However, if you have more than one cell in the table would get the same name, those cells will be referred to using the regular column-row addresses.

If the table has a header row but no header column you can do this:

Note that in this case "price" and "factor" refer to the entire column (see how the both columns are colored). For the given formula, Numbers will use only the cells in the same row as the formula (=A2*B2) but for something like =MAX(price) it will use the entire column.

Personally, I much prefer the standard column-row addressing. It is simpler, clearer, shorter, and makes it a lot easier to debug formulas. It also makes it a whole lot easier for those of us here trying to help if the person posting their problem does not use headers as labels.