Convert Number to Words LAMBDA (Very Large Numbers)

Occasional Contributor

I created this LAMBDA Function "Number_To_Words" in order to convert Numbers to Words (eg. 2813 can be written as Two Thousand Eight Hundred Thirteen in words)

 

First Parameter of the function is Number and second is Optional (By default Indian No. System, 2 for International No. System, 1 for Indian)

 

It can convert up to 100 Trillion (International) and Lakhs of Crores (Indian).

 

Avoided conversion after decimal for now but that can be implemented as well. If Number has decimal, function will Round it first and then give results.

 

YT Video Link - https://www.youtube.com/watch?v=u1gzAcwmlpo

 

Gist (GitHub) Link - https://gist.github.com/Bhavya2502/8413a0e6af783ad18e72419eca47ad09

 

What are your thoughts on this ?

 

@Peter Bartholomew  @Sergei Baklan  

10 Replies

@Bhavya250203 

That's interesting to compare with @Peter Bartholomew formula suggested couple of years ago https://techcommunity.microsoft.com/t5/excel-blog/announcing-let/ba-p/1233572

=LET(
   Denom, {" Trillion ";" Billion ";" Million ";" Thousand ";" Dollars ";" Cents"},
   Nums, {"","One","Two","Three","Four","Five","Six","Seven","Eight"," Nine"},
   Teens, {"Ten","Eleven","Twelve","Thirteen","Fourteen","Fifteen","Sixteen","Seventeen","Eighteen","Nineteen"},
   Tens, {"","Ten","Twenty"," Thirty"," Forty","Fifty","Sixty"," Seventy","Eighty","Ninety"},
   grp, {0;1;2;3;4;5},
   LET(
      N, SUBSTITUTE( TEXT( A1, REPT(0,15)&".00" ),".","0"),
      H, VALUE( MID( N, 3*grp+1, 1) ),
      T, VALUE( MID( N, 3*grp+2, 1) ),
      U, VALUE( MID( N, 3*grp+3, 1) ),
      Htxt, IF( H, INDEX( Nums, H+1 ) & " Hundred ", "" ),
      Ttxt,  IF( T>1, INDEX( Tens, T+1 ) & IF( U>0, "-", "" ), " " ),
      Utxt, IF( (T+U),  IF( T=1, INDEX( Teens, U+1 ), INDEX(Nums, U+1 ) ) ),
   CONCAT( IF( H+T+U, Htxt & Ttxt & Utxt & Denom, "" ) )
   )
)

Lot of changes in Excel from that time.

Will play with both some later.

@Sergei Baklan My goal was actually two fold for both Indian as well as International Number System. Both are very different in terms of place value of numbers. I first made for Indian and then adapted as per International System (With few addons, Some steps and Lookup Tables were common).

@Sergei Baklan 

Thank you for the reference.  Due to a clumsy transfer of data to a new computer, I had lost the original English version and was going to reconstruct it from a German language version I had written or the French version @Frederic LE GUEN had developed from the original.  The original workbook predates the CONCAT function and relied upon a multitude of defined names to conduct array calculations pre-dynamic-array.

 

Something that occurred when I tried to set up a new workbook from the published formula was that the formula refused to commit.  I tracked that back to the removal of '.' (period) as a valid character within a LET local name.

 

TextDollarAmountλ
= LAMBDA(target, 
    LET(
        Denom, {" Million, ";" Thousand ";" Dollars ";" Cents"},
        Nums,  {"","One","Two","Three","Four","Five","Six","Seven","Eight"," Nine"},
        Teens, {"Ten","Eleven","Twelve","Thirteen","Fourteen","Fifteen","Sixteen","Seventeen","Eighteen","Nineteen"},
        Tens,  {"","Ten","Twenty"," Thirty"," Forty","Fifty","Sixty"," Seventy","Eighty","Ninety"},
        grp,   {0;1;2;3},
        
        N,     SUBSTITUTE( TEXT( target, REPT(0,9)&".00" ),".","0"),
        H,     VALUE( MID( N, 3*grp+1, 1) ),
        T,     VALUE( MID( N, 3*grp+2, 1) ),
        U,     VALUE( MID( N, 3*grp+3, 1) ),
        Htxt,  IF( H, INDEX( Nums, H+1 ) & " Hundred ", "" ),
        Ttxt,  IF( T>1, INDEX( Tens, T+1 ) & IF( U>0, "-", "" ), " " ),
        Utxt,  IF( (T+U),  IF( T=1, INDEX( Teens, U+1 ), INDEX(Nums, U+1 ) ) ),
        CONCAT(IF(H+T+U, Htxt & Ttxt & Utxt & Denom, "") )
    )
 );

 

The good thing about turning the formula into a Lambda function is that it can be applied to many values within a workbook.

 

= MAP(target, TextDollarAmountλ)

 

 

@Bhavya250203 

Please accept my apologies for conducting an exchange with @Sergei Baklan on your thread.

I must admit, I wondered how you dealt with the more irregular patterns of Lakh and Crore (I find the number formatting to be enough of a challenge).  Presumably you could get some benefit from having  multiple groups of 2 digits rather than my 5 groups of 3 digits?

@Bhavya250203 

I have now downloaded your Lambda function.  Congratulations, there is nothing short in terms of ambition in that function!  I did try

= MAP(target, format, Indian.Number_To_Words)

where 'format' is a range containing {1;0;1} but it only returned the first number.  

[Added.  Only the first number had no decimal digits, so MAP works OK with your function]

The "Indian" prefix came about because I specified a Name space; I have never done that before.

The only thing I wondered about was whether it is necessary to be able to switch between international and local practice for reporting numbers?  Wouldn't two functions have done the job as well and permitted some further optimisation of the individual code?

 

 

 

Thanks a lot for the praise. Actually I originally made the function as per Indian Number System but then thought of adding more flexibility of choosing between the two systems as many steps were common in them.

@Peter Bartholomew , actually I took your initial formula in the comments to the blog here with LET() announcement . Removed dot in names, added "Billion" to Denom and expanded grp (from 0 to 5).

 

Yes, these days it's more reasonable to use it as custom function. By the way, UDF is reserved for VBA functions, will we use the same abbreviation for lambdas or one day another term appears?

@Bhavya250203 

I agree with @Peter Bartholomew , it's better to separate functions for each language. If only your are not going to generalize it and include few dozens of languages.

 

That's normal practice. For example, we don't have FILL() function with parameters Up and Down, we have FillUp() and FillDown().

May be we can call them Custom LAMBDA Function ??
Yes you are right Sir. Separating gives clarity as well