*Emphasis:
* generate a number "weighted" based on a sort order
any good?
*
Yes. Someone else made a suggestion that started my thinking. I think I
started off over-thinking this and making it too hard. <g> I think by in
the morning, I will have something workable that I can use, and not too
hard to do. Thanks for your thoughts!
**Thank you.
**Robert*
*Genealogy without documentation is mythology! Always SOURCE your work.*
On Sun, Jan 1, 2012 at 6:07 PM, Dr John C Bullas <john.bullas@gmail.com>wrote:
> **
>
>
> OK
>
> you could generate a false highly accurate number to enable you to do
> this sorting
>
> you want a number which would give you whatever search order you want
> based on the priority of the columns
>
> if you had a column next to each proper column which gave the order of
> entries for that column, 1,2,3,4,5,6,7 et seq convert this to
> 01,02,03,04,05,06,07
>
> do that for each column of values making sure you "expand" each one of
> the order entries is as "wide" as the largest number of digits for any
> of the columns, if there are 1000 in one column all the other columns
> are padded out so the one ablve is now 0001,0002,0003 et seq.
>
> now based on the sort ordering we now assemble long strings of text
> combining the ordering references in the order of the sort so
>
> colA / colB / colC/ colD/ colE
> 001 002 003 004 005
>
> B A C D E would return
>
> 002001003004005
>
> D E A B C would return
>
> 004005001002003
>
> would this weight an ordering of this long string as a number
> in the order you want
>
> generate a number "weighted" based on a sort order
>
> any good?
>
> On 1 January 2012 23:20, Robert E. Carneal
> <kentuckygenealogist@gmail.com> wrote:
> > *Emphasis:
> > * you could convert each letter to its letter number a=1 b=2 etc etc
> > convert that to binary and stuff in any spacer zeros so you could then
> grab
> > 5 decrypt it to the letter, grab the next five etc etc
> >
> > Or have I entirely lost the plot?
> >
> > *By jove, I think you might have it!!
> > That might work. You have given me some ideas. What I am trying to do is
> > sort some names on more than three columns to include the birth date, and
> > death date. I thought if I could convert that to a binary number (or
> > decimal number), I could convert that to a really high base number
> format,
> > and then sort on that value. Then everyone would be sorted correctly,
> and I
> > get around the three column sort limitation. I hope I made sense.
> >
> > **Thank you.
> >
> > **Robert*
> > *Genealogy without documentation is mythology! Always SOURCE your work.*
> >
> > * *
> > Or have I entirely lost the plot?
> >
> >
> > On Sun, Jan 1, 2012 at 4:51 PM, Dr John C Bullas <john.bullas@gmail.com
> >wrote:
> >
> >> **
> >>
> >>
> >> had a think
> >>
> >> like for like: you need something that can be reversed once
> >> "encrypted", the only way you could convert a name to a true binary
> >> number and then decode it would be to have a massive lookup list of
> >> binary numbers (serial numbers fo each name) and associated names and
> >> a way of ensuring each encryption was unique for different names and
> >> the same for each time the same name was "encoded"....
> >>
> >> I think what you want is encryption if it is to be reversible and
> >> someone has probably written a cypher and de-cypher script to do
> >> this.....
> >>
> >> letter to chunk: As you have a base 26 alphabet, trying to "reduce"
> >> this to a base 2 format means you could generate a binary number for
> >> each letter from 1 to 26 (1 to 11010 or 10000 to 11010 if you fill in
> >> spaces to generate strings of 5 digit phrases ) , so consider the name
> >> smith would expand five fold, decrypting this would be easy so you
> >> need a cypher and a like for like letter transposition can be cracked
> >> based on number or letter frequency
> >>
> >> you could convert each letter to its letter number a=1 b=2 etc etc
> >> convert that to binary and stuff in any spacer zeros so you could then
> >> grab 5 decrypt it to the letter, grab the next five etc etc
> >>
> >> Or have I entirely lost the plot?
> >>
> >> Dr B
> >>
> >> On 1 January 2012 21:48, Robert E. Carneal
> >> <kentuckygenealogist@gmail.com> wrote:
> >> >
> >> > *Is it possible to take a person's name, and convert it to a binary
> >> number?
> >> > I can convert any number, any number system to a binary number, but
> can't
> >> > seem to figure out a way to convert names similar to:
> >> >
> >> > DoeJohnWilliam
> >> > EsrankKathyMaria
> >> > MooreJamesAndrew
> >> > ThomasRaymondDavid
> >> >
> >> > And so on. When I looked on The Internet, it kept wanting me to
> convert
> >> > names to Soundex codes
>
[Non-text portions of this message have been removed]
Be sure to check out TechTrax Ezine for many, free Excel VBA articles! Go here: http://www.mousetrax.com/techtrax to enter the ezine, then search the ARCHIVES for EXCEL VBA.
----------------------------------
Visit our ExcelVBA group home page for more info and support files:
http://groups.yahoo.com/group/ExcelVBA
----------------------------------
More free tutorials and resources available at:
http://www.mousetrax.com
----------------------------------
Tidak ada komentar:
Posting Komentar