Using logicals in arrayfun

Hi,

I have two arrays:

tempA, size(10, 18)

tempB, size(1, 10) (is a column vector),

For each row in tempA, I want to extract the number of columns specified for that row by tempB. However, there are certain rows in tempB that are 'nan'.

u =  arrayfun(@(x,y) x{1}(1:y), tempA, tempB, 'UniformOutput', false);

Can I use logical in the arrayfun so that it automatically excludes cases that are 'nan'?

Thank you

Answer 1

[Walter Roberson]

u =  arrayfun(@(x,y) x{1}(1:max(0,y)), tempA, tempB, 'UniformOutput', false);

The secret here is that max(0,VALUE) is 0 if VALUE is nan. 1:0 is then empty, and indexing by empty is well-defined as being empty.

[Rashi Monga]

this was awesome! thanks

[Paul]

Missing a parenthesis after (0,y), but, that aside, u will contain cell elements that are empty arrays, if that's o.k. I thought those cases are to be excluded from the result.

[Walter Roberson]

Fixed the missing ) thanks!

Answer 2

[Paul]

rng(100)
tempA = rand(10,18);
tempB = 1:10;
u =  arrayfun(@(x,y) x{1}(1:y), num2cell(tempA,2), tempB.', 'UniformOutput', false)

u = 10x1 cell array
    {[                                                               0.5434]}
    {[                                                        0.2784 0.2092]}
    {[                                                 0.4245 0.1853 0.8176]}
    {[                                          0.8448 0.1084 0.3361 0.3819]}
    {[                                   0.0047 0.2197 0.1754 0.0365 0.2100]}
    {[                            0.1216 0.9786 0.3728 0.8904 0.5447 0.3402]}
    {[                     0.6707 0.8117 0.0057 0.9809 0.7691 0.1781 0.6023]}
    {[              0.8259 0.1719 0.2524 0.0599 0.2507 0.2377 0.3878 0.3404]}
    {[       0.1367 0.8162 0.7957 0.8905 0.2859 0.0449 0.3632 0.0921 0.3131]}
    {[0.5751 0.2741 0.0153 0.5769 0.8524 0.5054 0.2043 0.4635 0.6340 0.6576]}

If all the odd numbered indices of tempB are nan:

tempB(1:2:end) = NaN;
u =  arrayfun(@(x,y) x{1}(1:y), num2cell(tempA(~isnan(tempB),:),2), tempB(~isnan(tempB)).', 'UniformOutput', false)

u = 5x1 cell array
    {[                                                        0.2784 0.2092]}
    {[                                          0.8448 0.1084 0.3361 0.3819]}
    {[                            0.1216 0.9786 0.3728 0.8904 0.5447 0.3402]}
    {[              0.8259 0.1719 0.2524 0.0599 0.2507 0.2377 0.3878 0.3404]}
    {[0.5751 0.2741 0.0153 0.5769 0.8524 0.5054 0.2043 0.4635 0.6340 0.6576]}

[Rashi Monga]

thank you for your response! this was helpful. :)