This works: x = 1.2345 SET DECIMALS TO 4 SET FIXED ON ? TRANSFORM(x, "999.9999")

Laurie

On 31 July 2017 at 11:34, Laurie Alvey trukker41@gmail.com wrote:

SET DECIMALS TO n only works when SET FIXED is ON.

Laurie

On 30 July 2017 at 23:38, Gene Wirchenko genew@telus.net wrote:

...

Hello:

 I have just been bitten by SET DECIMALS.  (Is there really any use

for this besides making programming just a bit more difficult?)

 I have a table with values to four decimal places.  I use

<vfp> transform(thevalue,"999999.9999") </vfp> to create the string representation. With SET DECIMALS set to its default value, a value of 0.0123 is converted to " 0.0100" which loses two digits of precision.

 I had this problem with another data item that had more than two

decimal places. I wrote a special function to handle it by setting SET DECIMALS to the number of decimal places I needed, doing the transform(), and setting SET DECIMALS back to the default.

 Why did I do that?  Because the number of decimal places can vary

and I was generating a line of print in one line of code, BUT also, because I really do not understand the point of SET DECIMALS.

 Could someone please clue me in?

Sincerely,

Gene Wirchenko

[excessive quoting removed by server]

It makes a difference to the calculated value .... Only difference between two functions below is one sets decimal to 2 and other to 16. The 16 version produces expected results.

? DectoBase36_2(2176782335), DectoBase36(2176782335), "Expecting ZZZZZZ" && Returns "100000", "ZZZZZZ" ? DectoBase36_2(2176782334), DectoBase36(2176782334), "Expecting ZZZZZY" && Returns "100000", "ZZZZZY"

lnVal = 2176782335 SET DECIMALS TO 2 ? INT(LOG(m.lnVal) / LOG(36)), "Expecting 5" && Returns 6

SET DECIMALS TO 16 ? INT(LOG(m.lnVal) / LOG(36)), "Expecting 5" && Returns 5

FUNCTION DectoBase36

LPARAMETERS vnVal

LOCAL lnDecimals, lnPwr, lcString, lnMult, lnInt

STORE "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ" TO lcValues

lndecimals = SET("DECIMALS") SET DECIMALS TO 16

lnPwr = INT(LOG(m.vnVal) / LOG(36)) lcString = ""

FOR ln = m.lnPwr TO 0 STEP -1

lnMult = 36^m.ln lnInt = INT(m.vnVal / m.lnMult) vnVal = m.vnVal -(m.lnInt * lnMult)

lcString = m.lcString + SUBSTR(m.lcValues, m.lnInt + 1, 1)

ENDFOR

SET DECIMALS TO m.lndecimals

RETURN lcString

ENDFUNC

FUNCTION DectoBase36_2

LPARAMETERS vnVal

LOCAL lnDecimals, lnPwr, lcString, lnMult, lnInt

STORE "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ" TO lcValues

lndecimals = SET("DECIMALS") SET DECIMALS TO 2

lnPwr = INT(LOG(m.vnVal) / LOG(36)) lcString = ""

FOR ln = m.lnPwr TO 0 STEP -1

lnMult = 36^m.ln lnInt = INT(m.vnVal / m.lnMult) vnVal = m.vnVal -(m.lnInt * lnMult)

lcString = m.lcString + SUBSTR(m.lcValues, m.lnInt + 1, 1)

ENDFOR

SET DECIMALS TO m.lndecimals

RETURN lcString

ENDFUNC

-----Original Message----- From: ProfoxTech [mailto:profoxtech-bounces@leafe.com] On Behalf Of Ted Roche Sent: Monday, 31 July 2017 10:05 PM To: profoxtech@leafe.com Subject: Re: SET DECIMALS Gotcha

 I have just been bitten by SET DECIMALS.  (Is there really any 

use for this besides making programming just a bit more difficult?)

There may have been, when it was invented in the dBASE days. I believe it does not affect values, only their display.

 I have a table with values to four decimal places.  I use <vfp>
      transform(thevalue,"999999.9999") </vfp> to create the 

string representation. With SET DECIMALS set to its default value, a value of 0.0123 is converted to " 0.0100" which loses two digits of precision.

What if you multiplied the number by 10^4? Would the precision still be there, only not displayed?

 I had this problem with another data item that had more than two 

decimal places. I wrote a special function to handle it by setting SET DECIMALS to the number of decimal places I needed, doing the transform(), and setting SET DECIMALS back to the default.

 Why did I do that?

That is probably the crux of the matter.

BUT also, because I really do not understand the point of SET DECIMALS.

Is there someplace you could look it up? A reference guide of some sort?

-- Ted Roche Ted Roche & Associates, LLC http://www.tedroche.com

[excessive quoting removed by server]