Post has attachment

**Comma Separated Values**

I've put together some Excel VBA functions to do a constrained and highly predictable CSV format encode/decode. Given that this data is likely to be read way more often than written, the strategy is slow-write, fast-read. Also, keeping in mind Excel's 32k limit on strings, I've tried to compress the CSV as much as possible.

There's one thing missing in sdToCsv(): It needs code to remove trailing zeros in the fractional part of a number string. If anyone has some code to do that, it would be useful.

There's an Excel workbook with the code and some tests in it at

https://drive.google.com/open?id=0B2JnNP481pX4dTJFSlhscnFtbEE&authuser=0

The .bas module is at

https://drive.google.com/open?id=0B2JnNP481pX4cmlOQjFMRG1OYjg&authuser=0

Post has attachment

Marc,

I was playing with SDXL trying to recreate example of VOI calculation posted by Mike Middleton here www.mikemiddleton.com/Excel-Monte-Carlo-Simulation-with-Value-of-Information-2010.xlsx and described here www.mikemiddleton.com/Excel-Monte-Carlo-Simulation-with-Value-of-Information-2010.pdf

My results are published here: https://drive.google.com/folderview?id=0B463QavNC3_Mfk1YZFpVV2kyV3ZQNXEwS2hEVUphSUJsalozZWV1UlFrNWNKVFFOQld2bWs&usp=sharing

Couple of comments:

1) Ideas of how to pack the bracket values into the 1x100 array without VBA are welcome.

2) There's something strange going on with packing values into XML. Please, check out last flip. Integer array is packed with errors under the default setting. I guess 14 bits/word is only good for small values. Why can't I use more than 19 bits? "Float" packing produces significant aberrations even under highest bpw setting. I thought the errors should converge quicker.

I was playing with SDXL trying to recreate example of VOI calculation posted by Mike Middleton here www.mikemiddleton.com/Excel-Monte-Carlo-Simulation-with-Value-of-Information-2010.xlsx and described here www.mikemiddleton.com/Excel-Monte-Carlo-Simulation-with-Value-of-Information-2010.pdf

My results are published here: https://drive.google.com/folderview?id=0B463QavNC3_Mfk1YZFpVV2kyV3ZQNXEwS2hEVUphSUJsalozZWV1UlFrNWNKVFFOQld2bWs&usp=sharing

Couple of comments:

1) Ideas of how to pack the bracket values into the 1x100 array without VBA are welcome.

2) There's something strange going on with packing values into XML. Please, check out last flip. Integer array is packed with errors under the default setting. I guess 14 bits/word is only good for small values. Why can't I use more than 19 bits? "Float" packing produces significant aberrations even under highest bpw setting. I thought the errors should converge quicker.

Post has attachment

**SDXL Ver 1.1.1**

Thanks to Peter Kadar, I found and fixed a strange error in getStats. The Version 1.1.1 add-in is available at smpro.ca/SDXL.

**Update**It's now Version 1.1.2

Post has attachment

Might be worth considering this when creating DISTs in VBA

Thanks for the invite Marc. Look forward to learning lots from you and the other members

Post has attachment

A couple of updates:

Returns a permutation index that will select out the elements corresponding to TRUE (<>0) in the argument SIP. Essential to constructing a conditional probability distribution SIP.

I also extended

**sdFilterIndex(sip)**Returns a permutation index that will select out the elements corresponding to TRUE (<>0) in the argument SIP. Essential to constructing a conditional probability distribution SIP.

I also extended

**sdDiv()**and**sdSub()**to take a scalar first argument, in which case the second argument sets the result count.**sdSparkScat()**deletes the chart if np (third argument, number of points) is zero. Post has attachment

**Printed Ref Manual**

The printed edition of the Version 1.0 SDXL reference manual is now available from CreateSpace, Amazon to follow in a week or so.

Post has attachment

**SDXL Version 1.0**

Welcome to the new SDXL User Community. This should be a whole lot more convenient than a G+ Page. Please feel free to say hello and invite anyone you know who may be interested.

This version of SDXL mainly adds new graphic capabilities, including doing histogram/percentile charts the way I like them.

*New Modules:*

**Canvas**: An object that provides low level functions that use Excel’s Chart and Shapes objects directly.

**SmpSIP**: An object that does things with a SIP.

**phChart**: An object that helps to make a histogram and percentile curve chart.

**sdDrawSip**: Finally got the SIP histogram and percentile curve graphic the way I like it using the new charting classes.

**sdSparkScat**: Added a Sparkline styled scatter plot.

As usual, the reference manual, add-in and workbooks are at http://smpro.ca/SDXL

If you have any suggestions, questions,

*code*, ... post here, please.

Wait while more posts are being loaded