Adjusting table columns by a grouped value

When working with my dataset, I find that I need to adjust the data in ways that makes the plots more understandable. In this particular example, I'm trying to adjust the times in a table by finding the minimum value for every run, then subtracting the minimum value to every run so that each run in my dataset all start at t=0 on a plot.

I can find the mimium value, and have created a table with those values, however I am not sure how to apply these values to the main table, so that I can create an additional column with the adjusted time. I believe it has something to do with rowfun, but I can't figure out how to get around the different sizes of the rows to get a proper calculation. I will be adding what I have so far in an attachment. Thank you.

EDIT:

If it'll make any sense, I'm trying to do something akin to this in the code:

adjtfcn= @(times,mintime)(minus(times,mintime));

test.adjtimes=rowfun(adjtfcn,...

test, "inputvariables", [test.times,mint.min_times],...

"groupingvariables","cycleNumber")

Answer 1

[Star Strider]

One optiopn is to sue findgroups and then accumarray to create the different tables. They need to be separate since they are differnt lengths. The results appear to be sorted correctly by the ‘times’ variable.

Try this —

load('test')
whos('-file','test')

  Name      Size              Bytes  Class    Attributes

  mint       -                 4683  table              
  test       -              1015708  table              

mint

mint = 26x3 table
    cycleNumber    GroupCount    min_times 
    ___________    __________    __________

        0             7176           300.03
        1             7016            71966
        2             1667       1.4203e+05
        3             3263        1.759e+05
        4             3203       2.0843e+05
        5             3148       2.4037e+05
        6             3100       2.7176e+05
        7             4390       3.0266e+05
        8             1833       3.6557e+05
        9             1691       3.8382e+05
        10            1656       4.0064e+05
        11            1634       4.1711e+05
        12            1616       4.3336e+05
        13            5676       4.4942e+05
        14             534       5.0608e+05
        15             443       5.1133e+05

test

test = 59499x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

         0         300.03    3.6273
         0         300.03    3.6289
         0         310.03    3.6984
         0         320.03    3.7031
         0         330.03    3.7052
         0         340.03    3.7065
         0         350.03    3.7074
         0         360.03     3.708
         0         370.03    3.7084
         0         380.03    3.7085
         0         390.03    3.7088
         0         400.03    3.7089
         0         410.03    3.7089
         0         420.03     3.709
         0         430.03    3.7091
         0         440.03     3.709

[G,id] = findgroups(test.cycleNumber);

Result = accumarray(G, (1:numel(G)).', [], @(x){test(x,:)})

Result = 26x1 cell array
    {7176x3 table}
    {7016x3 table}
    {1667x3 table}
    {3263x3 table}
    {3203x3 table}
    {3148x3 table}
    {3100x3 table}
    {4390x3 table}
    {1833x3 table}
    {1691x3 table}
    {1656x3 table}
    {1634x3 table}
    {1616x3 table}
    {5676x3 table}
    { 534x3 table}
    { 443x3 table}

Result{1}

ans = 7176x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

         0         300.03    3.6273
         0         300.03    3.6289
         0         310.03    3.6984
         0         320.03    3.7031
         0         330.03    3.7052
         0         340.03    3.7065
         0         350.03    3.7074
         0         360.03     3.708
         0         370.03    3.7084
         0         380.03    3.7085
         0         390.03    3.7088
         0         400.03    3.7089
         0         410.03    3.7089
         0         420.03     3.709
         0         430.03    3.7091
         0         440.03     3.709

Result{end}

ans = 4374x3 table
    cycleNumber      times       EcellV
    ___________    __________    ______

        25          5.856e+05    3.8021
        25          5.856e+05    3.8025
        25         5.8561e+05    3.8104
        25         5.8562e+05    3.8122
        25         5.8563e+05    3.8134
        25         5.8564e+05    3.8143
        25         5.8565e+05    3.8151
        25         5.8566e+05    3.8159
        25         5.8567e+05    3.8165
        25         5.8568e+05     3.817
        25         5.8569e+05    3.8175
        25          5.857e+05     3.818
        25         5.8571e+05    3.8185
        25         5.8572e+05    3.8191
        25         5.8573e+05    3.8194
        25         5.8574e+05    3.8199

.

[Desmond]

So if I understand this correctly, this is making 26 different tables? If that's the case, how would I reference them all when making the plot? Also, if I'm understanding this correctly, even if I do this, I still have the issue where despite separating them into different tables, they still aren't corrected by the time value.

Perhaps I should explain it in a different way. Let's take Result{1} for instance. The minimum time value for that one is 300.03. I want to take all of the time values in Result{1}, and subtract them at 300.3, so that they start from 0. I want to do this for all of the groups, so when I plot them, they all start at 0. Currently, if I plot them, say using gscatter(), I end up getting them starting at when they started in the tests instead of 0.

That way, the data would look something like this:

cycleNumber times EcellV

___________ ______ ______

0 0 3.6273

0 0 3.6289

0 10.03 3.6984

[Star Strider]

That’s an easy fix. For every table, subtract the first time value from the resto f them.

load('test')
whos('-file','test')

  Name      Size              Bytes  Class    Attributes

  mint       -                 4683  table              
  test       -              1015708  table              

mint

mint = 26x3 table
    cycleNumber    GroupCount    min_times 
    ___________    __________    __________

        0             7176           300.03
        1             7016            71966
        2             1667       1.4203e+05
        3             3263        1.759e+05
        4             3203       2.0843e+05
        5             3148       2.4037e+05
        6             3100       2.7176e+05
        7             4390       3.0266e+05
        8             1833       3.6557e+05
        9             1691       3.8382e+05
        10            1656       4.0064e+05
        11            1634       4.1711e+05
        12            1616       4.3336e+05
        13            5676       4.4942e+05
        14             534       5.0608e+05
        15             443       5.1133e+05

test

test = 59499x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

         0         300.03    3.6273
         0         300.03    3.6289
         0         310.03    3.6984
         0         320.03    3.7031
         0         330.03    3.7052
         0         340.03    3.7065
         0         350.03    3.7074
         0         360.03     3.708
         0         370.03    3.7084
         0         380.03    3.7085
         0         390.03    3.7088
         0         400.03    3.7089
         0         410.03    3.7089
         0         420.03     3.709
         0         430.03    3.7091
         0         440.03     3.709

VN = test.Properties.VariableNames;

[G,id] = findgroups(test.cycleNumber);

Result = accumarray(G, (1:numel(G)).', [], @(x){test(x,:)})

Result = 26x1 cell array
    {7176x3 table}
    {7016x3 table}
    {1667x3 table}
    {3263x3 table}
    {3203x3 table}
    {3148x3 table}
    {3100x3 table}
    {4390x3 table}
    {1833x3 table}
    {1691x3 table}
    {1656x3 table}
    {1634x3 table}
    {1616x3 table}
    {5676x3 table}
    { 534x3 table}
    { 443x3 table}

for k = 1:numel(Result)
    Result{k}.times = Result{k}.times - Result{k}.times(1);
end

Result{1}

ans = 7176x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

         0              0    3.6273
         0          0.002    3.6289
         0         10.002    3.6984
         0         20.002    3.7031
         0         30.002    3.7052
         0         40.002    3.7065
         0         50.002    3.7074
         0         60.002     3.708
         0         70.002    3.7084
         0         80.002    3.7085
         0         90.002    3.7088
         0            100    3.7089
         0            110    3.7089
         0            120     3.709
         0            130    3.7091
         0            140     3.709

Result{end}

ans = 4374x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

        25              0    3.8021
        25          0.002    3.8025
        25         10.002    3.8104
        25         20.002    3.8122
        25         30.002    3.8134
        25         40.002    3.8143
        25         50.002    3.8151
        25         60.002    3.8159
        25         70.002    3.8165
        25         80.002     3.817
        25         90.002    3.8175
        25            100     3.818
        25            110    3.8185
        25            120    3.8191
        25            130    3.8194
        25            140    3.8199

figure
hold on
for k = 1:numel(Result)
    plot(Result{k}.times, Result{k}.EcellV, 'DisplayName',sprintf('Cycle # %2s',Result{k}.cycleNumber(1)))
end
hold off
grid
xlabel(VN{2})
ylabel(VN{3})
legend('Location','eastoutside')

Does this look correct?

.

[Desmond]

Yes that's exactly what I'm looking for. Thanks! Really the only thing I need to figure out left is the color coding, since matlab has so few options!

[Steven Lord]

There are three ways to specify the color of a line, as shown on this documentation page. There are only 8 that are named or have "short names" (red, green, blue, cyan, magenta, yellow, black, and white) but you can specify an RGB triplet or a hex color code. For example, if I want a line with a lot of blue, some red, and only a dash of green:

c = [0.4 0.1 0.9];
plot(1:10, 1:10, 'Color', c, 'LineWidth', 4) % Make it thick so it's easy to see

That's not a half bad purple. Or I could have used a hex color code. For a lot of red, some green, and only a little blue:

c = '#E08018';
figure
plot(1:10, 1:10, Color = c, LineWidth = 4)

I'd call that orange or maybe a dark mustard. [Naming colors can be tricky.]

[Star Strider]

As always, my pleasure!

Perhaps this —

load('test')
whos('-file','test')

  Name      Size              Bytes  Class    Attributes

  mint       -                 4683  table              
  test       -              1015708  table              

mint

mint = 26x3 table
    cycleNumber    GroupCount    min_times 
    ___________    __________    __________

        0             7176           300.03
        1             7016            71966
        2             1667       1.4203e+05
        3             3263        1.759e+05
        4             3203       2.0843e+05
        5             3148       2.4037e+05
        6             3100       2.7176e+05
        7             4390       3.0266e+05
        8             1833       3.6557e+05
        9             1691       3.8382e+05
        10            1656       4.0064e+05
        11            1634       4.1711e+05
        12            1616       4.3336e+05
        13            5676       4.4942e+05
        14             534       5.0608e+05
        15             443       5.1133e+05

test

test = 59499x3 table
    cycleNumber    times     EcellV
    ___________    ______    ______

         0         300.03    3.6273
         0         300.03    3.6289
         0         310.03    3.6984
         0         320.03    3.7031
         0         330.03    3.7052
         0         340.03    3.7065
         0         350.03    3.7074
         0         360.03     3.708
         0         370.03    3.7084
         0         380.03    3.7085
         0         390.03    3.7088
         0         400.03    3.7089
         0         410.03    3.7089
         0         420.03     3.709
         0         430.03    3.7091
         0         440.03     3.709

VN = test.Properties.VariableNames;

[G,id] = findgroups(test.cycleNumber);

Result = accumarray(G, (1:numel(G)).', [], @(x){test(x,:)});

for k = 1:numel(Result)
    Result{k}.times = Result{k}.times - Result{k}.times(1);
end

% Result{1}
% 
% Result{end}

cm = colormap(turbo(numel(Result)));

figure
hold on
for k = 1:numel(Result)
    plot(Result{k}.times, Result{k}.EcellV, 'DisplayName',sprintf('Cycle # %2s',Result{k}.cycleNumber(1)), 'Color',cm(k,:))
end
hold off
grid
xlabel(VN{2})
ylabel(VN{3})
legend('Location','eastoutside')

Choose whatever colormap you want.

.