Reading CSV as Structure Array

Dear MATLAB Community,

I am attempting to transform a CSV into a MATLAB structure for input into another function. At present I have come up with:

expression = table2struct(readtable('sample_Expression.csv'))
expression =
  136x1 struct array with fields:
    gene
    rawValue_1
    rawValue_2
    rawValue_3
    rawValue_4
    rawValue_5
    value

However, I am more interested in creating something like this:

expression =
  struct with fields:
        gene: {136x1 cell}
    rawValue: [136x5 double]
       value: [136x1 double]

What is the difference in terms of using the two different structures? Could someone suggest a way to modify my present script to create something in the second example?

Thanks for any help!

[Monika Jaskolka]

Can you attach your CSV file?

[Jonathan Josephs-Spaulding]

Hi Monika,

Here is a CSV file for your reference, thanks for any help or suggestions!

Answer 1

[Stephen23]

T = readtable('Test_Expression.csv', 'Decimal',',', 'Delimiter',';')

T = 122×7 table
                gene                rawValue_1    rawValue_2    rawValue_3    rawValue_4    rawValue_5    value
    ____________________________    __________    __________    __________    __________    __________    _____

    {'Chromosome_336:2795'     }        258           23            1            184            61        105.4
    {'Chromosome_2800:3723'    }        175            8            0             65            17           53
    {'Chromosome_3733:5016'    }        230            7            0             68            14         63.8
    {'Chromosome_22934:23872'  }        131           39            0             49            13         46.4
    {'Chromosome_23918:26731'  }        600          170            1            168            37        195.2
    {'Chromosome_30266:31084'  }         85            2            0             11             1         19.8
    {'Chromosome_99529:100836' }        486          117            0             96            29        145.6
    {'Chromosome_103207:104658'}        391           95            1             72            25        116.8
    {'Chromosome_108999:109910'}       1306          246            4            381           139        415.2
    {'Chromosome_196798:197514'}        159           65            0             78            21         64.6
    {'Chromosome_198322:199509'}         86           34            0             40             2         32.4
    {'Chromosome_200478:201332'}        695          253            2            202            54        241.2
    {'Chromosome_205772:206794'}        653          124            0            101            43        184.2
    {'Chromosome_207361:208146'}        456          100            0             85            31        134.4
    {'Chromosome_233915:234715'}         11            7            0              6             0          4.8
    {'Chromosome_293697:294947'}         50            3            0              1             0         10.8

S = struct('gene',{T.gene}, 'value',T.value)

S = struct with fields:
     gene: {122×1 cell}
    value: [122×1 double]

S.rawValue = [T.rawValue_1, T.rawValue_2, T.rawValue_3, T.rawValue_4]

S = struct with fields:
        gene: {122×1 cell}
       value: [122×1 double]
    rawValue: [122×4 double]

[Jonathan Josephs-Spaulding]

Thank you for this good start. However when I use the last line of code, the output is:

Scalar structure required for this assignment.

[Stephen23]

@Jonathan Josephs-Spaulding: so far you did not mention that your file is actually semi-colon separated with decimal commas. When you do not provide complete, accurate descriptions of the data, then it is expected that getting a working solution is unlikely.

I fixed my answer using the file that you uploaded here.

[Jonathan Josephs-Spaulding]

@Stephen Cobeldick thanks for your suggestion, I will try to keep best practice in the future if I post again. Thanks again for your help.

[Jonathan Josephs-Spaulding]

Hey @Stephen Cobeldick, when I include

% 'Decimal',',',

while reading the table in, I receive the error:

Error using readtable (line 216)
Invalid parameter name: Decimal.

However when this is removed, the answer that you provided works fine. The finally structure still remains as,

S =
  struct with fields:
        gene: {122x1 cell}
       value: [122x1 double]
    rawValue: [122x4 double]

Is there a difference here that I should be concerned with? Thanks for your advice!

[Stephen23]

"Is there a difference here that I should be concerned with?"

That depends on your data file: what decimal radix sign does your file actually use?

Did you actually look at the data values (not just at the "122x1 double"), to check if the complete decimal fraction was imported correctly? If the radix symbol does not match the file format, then it is possible that you will only get the integer portion (or something similar, I have not tried this).

Tip: do NOT open and save CSV files (or any text files) using Excel, unless you want to change the format and data of your files in unpredictable ways:

https://www.theregister.com/2020/08/06/excel_gene_names/

I answer a lot of questions on this forum where the user is not aware that they are working with two different file formats: one saved by some measurement device, another after they open&save the file using Excel.