How can I compress a table

I have an imported csv file. Containing a lot of rows How can I compress them so that al of the data of one year is stored in one row. (as in the picture below? I could not get the script from an much older matlab version running)

and what kind of type are the columns with the icon that looks like three sticky notes?

Kind regards Ellen

[Christopher McCausland]

Hi Ellen,

The best way forward is probably to share a few rows of your data. Could you also explain a little more how you are expecing the output to look like? Sharing the old code may help here too.

In terms of the three sticky notes, I beleive they are known as 'pages', all that is being symbolised is that you have multiple data points stored in one structure field and row i.e. a 8784x1 double rather than a signal value as in the year feild. It makes a bit more sense here: Page-wise matrix multiplication - MATLAB pagemtimes - MathWorks Switzerland read the "More About" for a nice diagrame.

Hope this helps,

Christopher

[Ellen]

This is the file and I found out how to create the first row with years.

But if i have the table put in my workspace how can I tell that I want to read te rows.

• read_knmivliss.m is a copy of the original file but that uses de csv directly from the source.

I have changed and used so far:

%% read KNMI data station Vlissingen:

% Gedownload van http://projects.knmi.nl/klimatologie/uurgegevens/selectie.cgi

% Geselecteerd zijn: wind- en luchtdrukgegevens

% Coordinaten en terreinhoogte uit deze datafile maar

% gegevens meethoogte vanuit KNMI, luchtdruk op 1.5m +8m NNAP

meteo.data=struct('year', num2cell(1996:2022));

meteo.source='KNMI https://www.knmi.nl/nederland-nu/klimatologie/uurgegevens';

meteo.location='310 = Vlissingen 51.44144N 3.59582E, terreinhoogte: 8.0 m NAP. Meethoogte 20 meter +nap';

meteo.units='wind in m/s, direction in degrees east of north, pressure in Pa, time in UTC';

%% read KNMI data:

noy=size(meteo.data,2); % number of years

%for indecade=0:2 % files span of 3 decades

%%I need to find de code to my own as table imported csv file which is given the name KNMI31019962023

I want the in de cell in the rows the corresponding data.It has tot look similar to te picture in de first post. Should I call the file meteo.data? But when I rename the data disappear.

Kind regards Ellen

Answer 1

[Stephen23]

"How can I compress them so that al of the data of one year is stored in one row. (as in the picture below?"

Your screenshot shows a non-scalar structure with fields 'year', 'month', etc.:

https://www.mathworks.com/help/matlab/matlab_prog/access-multiple-elements-of-a-nonscalar-struct-array.html

We can get such a structure by importing as a table and then a few simple manipulations on it:

T = readtable('KNMI310_1996-2023.csv')

T = 43848×7 table
    YYYY    MM    dd    HH    DD     FF      P  
    ____    __    __    __    ___    __    _____

    2008    1     1      1    120    30    10272
    2008    1     1      2    170    20    10271
    2008    1     1      3    130    10    10266
    2008    1     1      4    220    20    10261
    2008    1     1      5    200    10    10260
    2008    1     1      6    110    30    10257
    2008    1     1      7    150    20    10253
    2008    1     1      8     90    30    10252
    2008    1     1      9    100    30    10252
    2008    1     1     10    100    30    10252
    2008    1     1     11    110    50    10249
    2008    1     1     12    100    40    10245
    2008    1     1     13    100    50    10240
    2008    1     1     14    110    60    10235
    2008    1     1     15    110    60    10236
    2008    1     1     16    100    50    10234

U = varfun(@(v){v},T, 'GroupingVariables','YYYY')

U = 5×8 table
    YYYY    GroupCount        Fun_MM             Fun_dd             Fun_HH             Fun_DD             Fun_FF              Fun_P     
    ____    __________    _______________    _______________    _______________    _______________    _______________    _______________

    2008       8784       {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}
    2009       8760       {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}
    2010       8760       {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}
    2011       8760       {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}    {8760×1 double}
    2012       8784       {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}    {8784×1 double}

S = table2struct(U) % take a look at S in the variable viewer.

S = 5×1 struct array with fields:
    YYYY
    GroupCount
    Fun_MM
    Fun_dd
    Fun_HH
    Fun_DD
    Fun_FF
    Fun_P

Answer 2

[Peter Perkins]

Here's another possibility, using rowfun, that returns a table containing tables, which maybe is easier to use than what you asked for:

T = readtable("https://www.mathworks.com/matlabcentral/answers/uploaded_files/1554527/KNMI310_1996-2023.csv")

T = 43848×7 table
    YYYY    MM    dd    HH    DD     FF      P  
    ____    __    __    __    ___    __    _____

    2008    1     1      1    120    30    10272
    2008    1     1      2    170    20    10271
    2008    1     1      3    130    10    10266
    2008    1     1      4    220    20    10261
    2008    1     1      5    200    10    10260
    2008    1     1      6    110    30    10257
    2008    1     1      7    150    20    10253
    2008    1     1      8     90    30    10252
    2008    1     1      9    100    30    10252
    2008    1     1     10    100    30    10252
    2008    1     1     11    110    50    10249
    2008    1     1     12    100    40    10245
    2008    1     1     13    100    50    10240
    2008    1     1     14    110    60    10235
    2008    1     1     15    110    60    10236
    2008    1     1     16    100    50    10234

U = rowfun(@(MM,dd,HH,DD,FF,P){table(MM,dd,HH,DD,FF,P)},T, GroupingVariables="YYYY",OutputVariableNames="Data")

U = 5×3 table
    YYYY    GroupCount         Data     
    ____    __________    ______________

    2008       8784       {8784×6 table}
    2009       8760       {8760×6 table}
    2010       8760       {8760×6 table}
    2011       8760       {8760×6 table}
    2012       8784       {8784×6 table}

U.Data{1}

ans = 8784×6 table
    MM    dd    HH    DD     FF      P  
    __    __    __    ___    __    _____

    1     1      1    120    30    10272
    1     1      2    170    20    10271
    1     1      3    130    10    10266
    1     1      4    220    20    10261
    1     1      5    200    10    10260
    1     1      6    110    30    10257
    1     1      7    150    20    10253
    1     1      8     90    30    10252
    1     1      9    100    30    10252
    1     1     10    100    30    10252
    1     1     11    110    50    10249
    1     1     12    100    40    10245
    1     1     13    100    50    10240
    1     1     14    110    60    10235
    1     1     15    110    60    10236
    1     1     16    100    50    10234

U.Data{1}.DD(1:5) % etc.

ans = 5×1

120 170 130 220 200

[Ellen]

thanks!

Ellen