# How can I do this : {    [8:23] [24:39] [40:55] } --- these number are just an example

1 view (last 30 days)
Hello,
I have two arrays a = [1 2 3 4 5 5] and b = [6 7 5 8 9]. I need to write them in CAN bit packing block with these Format
c = {    [8:23] [24:39] [40:55] } ( Number are just an example).
c is the combination between a and b. I've tried with this code :
c = unique([a b]);
strc = num2str(c);
splitc = strsplit(strc);
joinc = strjoin(splitc, ':');
strjoin ={joinc};
and it was wrong . At least how can I add a delimiter such as '[' and ']' in one element in my array? e.g. a = [1 2 3]
a1 =--( it has to be string). Thanks for the help! :)

Show 1 older comment
jgg on 22 Dec 2015
I think the best thing to do would be to show us what the output should look like for your example. If I understand your problem correctly it should be {[1:9]}, correct?
CAN bit packing http://de.mathworks.com/help/xpc/io_ref/canbitpacking.html?s_tid=srchtitle I've got the number from a struct in workspace. c has to be cell array of strings.
Yes, that's correct , e.g. {[1:9]} . Maybe you could take a look in this link http://de.mathworks.com/help/xpc/io_ref/canbitpacking.html?s_tid=srchtitle

Renato Agurto on 22 Dec 2015
Do you want something like this?
['[' strc(1) ':' strc(4) ']']

yeah, can you tell me how to implements that stuff in this case : a1 = [0 32]; c1 = [31 63]; Result that I attempt : { [0:31] [32:63]} --- This one is string. Thanks in advance :D
one thing, It should be automatic cause the values of a1 and c1 could varies ,e.g. a1 = [0 8 16]; c1 =[7 8 16]; then the result should be : {[0:7]  }

Guillaume on 22 Dec 2015
Edited: Guillaume on 22 Dec 2015
First, learn to read the documentation. The documentation that you linked clearly says that: "The data type [...] must be a MATLAB® cell array vector. [...] The cell array elements must be of type double array". There are no strings involved in this.
Secondly, I'm not even sure you've understood what the can bit packing function does. The bit-pattern is not the same at all as the data you want to transmit. There are many bit patterns that could be valid for transmitting a = [1 2 3 4 5 5]. The following bit-pattern is a possiblity: c = {0, 1:2, 3:4, 5:7, 8:10, 11:13)}, that's the minimum pattern to transmit a but will not transmit b = [2 2 3 4 5 5] as there is not enough bits to transmit the first number. Equally valid pattern is: c = {0:9, 10:19, 20:30, 31:41, 42:52, 53:63}.
The bit-pattern is defined by:
• how many numbers you want to pack into a double
• the maximum value of each of these numbers
Given the maximum values you want to transmit
maxvalues = [2 2 6 4 10 8]; %for example.
You can calculate the smallest bit-pattern with:
bitsrequired = floor(log2(maxvalues)) + 1;
endbits = cumsum(bitsrequired) - 1;
assert(endbits(end) <= 63, 'more bits are required than can fit into a double');
bitpattern = arrayfun(@(s, e) (s:e), [0, endbits(1:end-1)+1], endbits, 'UniformOutput', false)

#### 1 Comment

Guillaume on 22 Dec 2015
Following onto the comment you wrote to Renato's answer while I was typing mine, if given inputs
startbits = [0 8 16];
endbits = [7 8 16];
To generate the corresponding bit pattern, you'd simply do:
bitpattern = arrayfun(@(s, e) s:e, startbits, endbits, 'UniformOutput', false);