Alternative to bitshift(A,k,N) as in 2014a , it gives a warning and I would like to avoid using it. But cannot understand alternative suggested.

I have a binary number

 % code
a = uint8(88) , b = dec2bin(a) --> b = 1011000
I would like to just get consecutive 11 without using bitshift(a,-3,2).
As in 2014a it gives following warning: Warning: BITSHIFT(A,K,N) will not accept integer valued input N in a future release. Use BITSHIFT(A,K,ASSUMEDTYPE) instead.

I cannot understand how to use bitshift(A,K,Assumedtype) for this issue.

Any idea how to get away without warning so that I am okay for newer versions.

Thanks for your help. Nirav

Answer 1

[Ashish Gudla]

If you need to truncate and return only 2 bits after shifting you can "bitand" the result with appropriate number.

for last bit, you use 00000001 --> 1

for last 2 bits, 00000011 --> 3

for last 3 bits, 00000111 --> 7

and so on

For example to get last 2 bits after right shifting 3 bits

bitand( bitshift(a,-3,'uint8') , 3)

[Nirav]

Thanks for the answer. Also, it seems to me that there is no added time benefit between

1) bitand(bitshift(a,-3),2^2-1) OR
2) bitand(bitshift(a,-3),'uint8'),2^2-1)

ASSUMEDTYPE is useful only if the input 'a' is of datatype double and NOT integers as for integers ASSUMEDTYPE is same datatype as input.

Thanks, Nirav