In a Google Sheet, I have 3 raw data worksheets within one google spreadsheet ('Category1'),('Category2'),('Category3'). These worksheets are constantly updated by people in my business but unfortunately the data isn't in a normailised form to be able to run efficient queries.
I would like to create a script that automatically generates a normailised Output ('Category1 Output'),('Category2 Output'),('Category3 Output') of this raw information that automatically updates itself when someone makes a change in the raw tabs.
In the google sheet below, I have provided an example of what one Category needs to look like. 'Category1' worksheet is the raw that is constantly updated by everyone. 'Category1Output' is the final output worksheet that automatically updates itself when an edit is made in the 'Category1' worksheet.
The Questioner has three sheets in a defined columnar layout - essentially several rows per data set and several columns (one for each time period). These sheets are not being replaced, but a summarised version was sought where filtering could be used to focus effectively on relevant data. Thus, each sheet to be translated from columnar format to row-wise format.
The process itself is straightforward. The source data comprised 64 products @ 8 data rows per product. The output records were @1,350.
The questioner's code was hung up on the conversion of data to the output format. The use of 8 rows of data per product is important, and the code includes a check that the quotient the total number of rows of data divided by eight is an integer. In addition, the names of the Source sheet and the Output sheet are called by name (
getSheetByName
) so that the code can be easily applied to any named Input Sheet and any named Output sheet. The only proviso is that both sheets must exists beforehand.Initial resolution of the questioner's code hiccup was successful and using the methodology of
getDataRange
andgetValues
before the loop greatly improved performance. There are two loops; one with a vertical orientation, moving through the rows of data; and the second with a horizontal orientation moving through the time-related columns. However performance was initially very inefficient and the code was timing out before completion.I modified the code to build a single 2D array and save it to the Output sheet just once at the end of the loops. This had a dramatic effect on performance. Total time to complete dropped from several minutes to less than 5 seconds.
UPDATE
The second element of the Questioner's code deals with updating data to an "Output Sheet" as changes are made to "Category" sheets. The questioner's code for the update was OK, but was missing the translation of the source range on the Category Sheet to establishing the target range on the Output sheet.
The solution involves a rubric based on mathematical number sequence. In this case, the mathematical sequence is the row numbers for products on the source sheet; each product occupies 8 rows, and the first row is #10, so the sequence is 10,18,26,34....
onEdit returns the range of the changed cell and
getRow
andgetColumn
can be used to establish the co-ordinates of the changed cell. The challenge is, knowing the actual row number that was changed, to establish what number in the sequence of rows (and hence the product name) that the actual row number represents. It is also extremely unlikely (eight to one) that the changed row will coincide with the first row for a product number.So it is necessary to apply the algorithm for mathematical sequences - twice. The formula to determine the nth number in a sequence of numbers is An = A1 + (D x(n-1)), where A1 is number for the first row (in our case, 10), D= the difference between each number in the sequence (in our case, 8), and n= the number in the sequence (in our case, the changed row number).
The first pass is to establish the position number in the sequence of numbers(product groups) represented by the actual changed row. It is very likely that the outcome is not an integer, i.e. it does not coincide with the first row for a Product group. So, the result is rounded down to the nearest integer, and the algorithm is processed a second time.
However this time we know the position of the sequence number, and we solve to find the value of the number. In this case, the formula is ((An-A1)/D)+1. This will return the row number in the Source sheet corresponding to the first row for the relevant Product group. We use this to identify what type of field was changed (Category, Display, etc).
The column number indicates the week number. Week 1 starts in Column F, so
get column
enables us to establish whether the change took place in a week column (or whether it was to left of Column F). If to the left, then "not my problem", if in F or higher, then it needs to be noted.Lastly, we do a
getRangeValue
for the Target sheet and look for a match of the week number in Column A AND the truncated Product name in column B. This provides the co-ordinates tosetValue
for the new value tracked from OnEdit.