Comments on: Data Table Basics

By: Dick Kusleika

Dick Kusleika — Fri, 14 Jan 2011 17:54:00 +0000

Apparently not well enough. I fixed the link to the screenshot, so that should be working now. As for updating posts from 2004 to be Excel 2010 compliant… I guess I could issue a disclaimer that six year old posts may not work on six month old software, but you’re the first to complain about it.

By: Andronicus

Andronicus — Fri, 14 Jan 2011 03:08:00 +0000

Well, that’s completely wrong for Excel 2010. BTW, your screen shot is blank. Did you proof this ? Pretty poor presentation, dude.

By: Robin Hammond

Robin Hammond — Wed, 13 Oct 2010 03:44:00 +0000

While appealing in small simple models with very short calculation chains, data tables really do slow models down dramatically. In many cases, the tables we see are performing a relatively simple sensitivity analysis, with one or two inputs from the rest of the model, but each scenario takes a full recalculation of the precedent chain. Two solutions:- don’t use them in complex model; or move the table to another spreadsheet that links back to the precedents using named ranges – this can easily be opened, updated, and calculated in isolation.

By: Peter

Peter — Mon, 11 Oct 2010 16:40:00 +0000

I can't seem to see what the advantage is over doing the calculation with proper cell referencing (absolute, relative or mixed) and then copy paste......

By: David Tanner

David Tanner — Sun, 06 Jan 2008 22:21:00 +0000

Greetings - I have forgotten how to do something in excel. I want to show the effects of changing an assumption on a range of resulting data. Specifically I have a key variable that drives revenue and I want to show the result in a 5 year P&L based on various values of the key variable
Key
Variable Year 1 Year 2 Year 3 Year 4 Year f
100
125
150
175
200 as so on

I belive there is also a table where you can have a changed various on a column and row and in the result quadrant underneath have a single result cell displayed that correlates to both.

Any help you can give on these two tables would be very much appreciated or directing me to a source.

Thanks so much

By: Jon Peltier

Jon Peltier — Tue, 15 Aug 2006 12:46:00 +0000

For a comprehensive explanation and example of data tables, look at this page on Dermot Balson's Modelling Excel web site:

http://www.westnet.net.au/balson/ModellingExcel/LookMaNoCode.shtml

I'm finally convinced.

By: Lori

Lori — Mon, 03 Apr 2006 14:26:00 +0000

Example above with tabs replaced by semicolons:
----

cost;Year1;Year2;1;internal rate of return
;;;2;=IRR(INDEX(A:C,D2,))
-20000;11000;12000;3;=TABLE(,D2)
-30000;15000;18000;4;=TABLE(,D2)

By: Lori

Lori — Mon, 03 Apr 2006 14:16:00 +0000

One way to use multiple input values in a data table is to insert a column index
and then use a lookup function in the table formula.

The example below calculates the IRR for three input columns:

ABCDE
-------------------------------------
costYear1Year21internal rate of return
2=IRR(INDEX(A:C,D2,))
-2000011000120003=TABLE(,D2)
-3000015000180004=TABLE(,D2)
...

The table range is D2:E4 with D2 as column input cell. Column D and row 2 can be hidden.

In this simple exampleno data table is needed, of course, but the method extends to more complex models.

By: Tushar Mehta

Tushar Mehta — Mon, 03 Apr 2006 03:47:00 +0000

Yes, data tables can be a very powerful construct. In fact, if one looks at how they are implemented it will quickly become apparent that MS broke its own rules about what a function is allowed to do. Once you specify that a range contains a table, you will notice that the cells in that range contain an array formula, TABLE(). Of course, the implementation of a data table requires that XL change the value of one (or two) cells in the worksheet, recalc the worksheet and then update the value in the data table cell. Try doing that with a regular function (or a UDF)!

For my doctorate, I used data tables extensively for analysis where algebraic analysis fell short. For example, in a real options model to evaluate a monopolist's investment opportunity, the first order derivative took a *complete* 8-1/2 x 11 page to write out. There was no way I could possibly analyze the equation and reach any meaningful conclusion. So, I resorted to numerical analysis and data tables.

The analysis quickly ran into various limits of the data table. One of the capabilities required was to run a macro before computing the value for each cell in the table. Basically, this would allow one to carry out an optimization as part of the table calculation. Another very useful capability would be the ability to specify a vector of values, not just a single row (or column value). Basically, I wanted not just r1, r2, etc. going down the rows in a single column but a vector such as (a1,b1), (a2, b2), etc. going down the rows in 2 (or more) columns.

The result was a homegrown add-in called SuperTable. After completing the dissertation I figured I'd make it available to the world at large. In the process of cleaning it up and making it presentable, I also enhanced it to be more flexible. I also tried to figure out how to make it work like XL's TABLE() function. That was a mistake. While it still works, it's very, very difficult to document. So, I put it aside.

Every so often, I open the file and wonder what I can do to clean it up and post on my website. Maybe, some day...

By: Erik Protzman

Erik Protzman — Sun, 02 Apr 2006 06:50:00 +0000

If anyone can come up with a method for cleanly linking a data table to a constant on another sheet I will be most grateful. I want to link the data table row and column inputs to constants on another sheet. Is this possible in some obscure way? Or has anyone successfully reproduced the data table tool in VBA? Could please shed some light? Thanks in advance.