Trouble using container with cell type for key and structure array for values..

Given this table,

TestTable_valid =

11×6 table

SerialNumber        ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
____________________    __________________    ___________________    __________    _________    _______________

{'A017441206210001'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
{'A017441229220004'}         {'7.0A'}              {'8.6A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
{'A020621202230003'}         {'0A'  }              {'7.7A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621203230003'}         {'0A'  }              {'7.6A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621207230002'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621208230001'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621210230006'}         {'1.2A'}              {'0.7' }          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621215230003'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020621222230004'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
{'A020622108240013'}         {'7.1A'}              {'0A'  }          {'Passed'}    {'VA2CE'}        {'Yes'}    
{'A020622109240009'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2CE'}        {'Yes'}   

I'm trying to create a container that uses the serial number as the key, and returns the other columns:

serials = TestTable_valid.SerialNumber
dataValues = TestTable_valid(:, 2:6)
dataStructArray = table2struct(dataValues)
% Create map: key = serial number, value = table row with C–G
lookupMap = containers.Map(serials, dataStructArray, 'UniformValues', false);  % needed because data types in structure may be different.

however, I get the following error:

Error using containers.Map

Specified value type does not match the type expected for this container.

Don't know what else to try...I'm not sure cell vs. character array, vs cell array of characters, so perhaps that is the issue.....

I appreciate the help...

Jorge

[Matt J]

Given this table,...

Are we? I see no .mat attachment.

[Walter Roberson]

It is not clear that containers.Map supports struct arrays for valueset

[Cris LaPierre]

Adding a runnable example.

TestTable_valid = readtable('TestTable_valid.xlsx')

TestTable_valid = 11x6 table
        SerialNumber        ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
    ____________________    __________________    ___________________    __________    _________    _______________

    {'A017441206210001'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A017441229220004'}         {'7.0A'}              {'8.6A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A020621202230003'}         {'0A'  }              {'7.7A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621203230003'}         {'0A'  }              {'7.6A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621207230002'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621208230001'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621210230006'}         {'1.2A'}              {'0.7' }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621215230003'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621222230004'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020622108240013'}         {'7.1A'}              {'0A'  }          {'Passed'}    {'VA2CE'}        {'Yes'}    
    {'A020622109240009'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2CE'}        {'Yes'}    

serials = TestTable_valid.SerialNumber

serials = 11x1 cell array
    {'A017441206210001'}
    {'A017441229220004'}
    {'A020621202230003'}
    {'A020621203230003'}
    {'A020621207230002'}
    {'A020621208230001'}
    {'A020621210230006'}
    {'A020621215230003'}
    {'A020621222230004'}
    {'A020622108240013'}
    {'A020622109240009'}

dataValues = TestTable_valid(:, 2:6)

dataValues = 11x5 table
    ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
    __________________    ___________________    __________    _________    _______________

         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
         {'7.0A'}              {'8.6A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
         {'0A'  }              {'7.7A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'0A'  }              {'7.6A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'1.2A'}              {'0.7' }          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
         {'7.1A'}              {'0A'  }          {'Passed'}    {'VA2CE'}        {'Yes'}    
         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2CE'}        {'Yes'}    

dataStructArray = table2struct(dataValues)

dataStructArray = 11x1 struct array with fields:
    ACCurrentPreSyncEn
    ACCurrentPostSyncEn
    PassFail
    Model
    LCRSwept_Yes_No

% Create map: key = serial number, value = table row with C–G
lookupMap = containers.Map(serials, dataStructArray, 'UniformValues', false);  % needed because data types in structure may be different.

Error using containers.Map
Specified value type does not match the type expected for this container.

[Jorge]

@Cris LaPierre, thanks for creating the table so that it can be imported and run with the code provided. Now I know to do this in the future!

If i don't convert the dataValues to a structure, then I get this other error:

Error using containers.Map
The number of keys and values must be the same.

So, all I want to be able to do is to use the SN (which comes from a different source) as a lookup into this table and extract the other fields. I guess I can do a row by row string compare and use the index of the matching row to extract the field values from the table...but I thought this would be the more elegant way....

[Walter Roberson]

As of R2025a (currently pre-release), you will be able to use

dataStructArray = num2cell(dataValues,2);

to create a cell array in which each row is a cell array containing the variables.

Unfortunately in R2024b and earlier, building such a cell array is most easily done using a loop.

[Jorge]

this worked....

TestTable_valid = readtable('TestTable_valid.xlsx')

TestTable_valid = 11x6 table
        SerialNumber        ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
    ____________________    __________________    ___________________    __________    _________    _______________

    {'A017441206210001'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A017441229220004'}         {'7.0A'}              {'8.6A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A020621202230003'}         {'0A'  }              {'7.7A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621203230003'}         {'0A'  }              {'7.6A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621207230002'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621208230001'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621210230006'}         {'1.2A'}              {'0.7' }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621215230003'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621222230004'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020622108240013'}         {'7.1A'}              {'0A'  }          {'Passed'}    {'VA2CE'}        {'Yes'}    
    {'A020622109240009'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2CE'}        {'Yes'}    

serials = cell(TestTable_valid.SerialNumber)

serials = 11x1 cell array
    {'A017441206210001'}
    {'A017441229220004'}
    {'A020621202230003'}
    {'A020621203230003'}
    {'A020621207230002'}
    {'A020621208230001'}
    {'A020621210230006'}
    {'A020621215230003'}
    {'A020621222230004'}
    {'A020622108240013'}
    {'A020622109240009'}

dataCell = table2cell(TestTable_valid(:, 2:6));
rowCells = mat2cell(dataCell, ones(size(dataCell,1),1),size(dataCell,2))

rowCells = 11x1 cell array
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}
    {1x5 cell}

lookupMap = containers.Map(serials, rowCells)

lookupMap = 
  Map with properties:

        Count: 11
      KeyType: char
    ValueType: any

fields = lookupMap(serials{1})

fields = 1x5 cell array
    {'7.1A'}    {'7.1A'}    {'Passed'}    {'VA2C'}    {'Yes'}

fprintf('SN: %s | Pre: %s | Post: %s | Result: %s | Model: %s | LCR: %s\n', ...
        serials{1}, fields{1}, fields{2}, fields{3}, fields{4}, fields

SN: A017441206210001 | Pre: 7.1A | Post: 7.1A | Result: Passed | Model: VA2C | LCR: Yes

Thank you for your help.

Answer 1

[Walter Roberson]

struct values are not permitted.

keys = 100:110;
values = struct('abc', num2cell(1:11))

values = 1x11 struct array with fields:
    abc

containers.Map(keys, values)

Error using containers.Map
Specified value type does not match the type expected for this container.

Answer 2

[Matt J]

Perhaps use a dictionary, instead of containers.Map?

keys =  compose("A0%.2d" , 1:15);
values = struct('abc', num2cell(1:15));

d=dictionary(keys, values),

d =

  dictionary (string --> struct) with 15 entries:

    "A001" --> 1x1 struct
    "A002" --> 1x1 struct
    "A003" --> 1x1 struct
    "A004" --> 1x1 struct
    "A005" --> 1x1 struct
    "A006" --> 1x1 struct
    "A007" --> 1x1 struct
    "A008" --> 1x1 struct
    "A009" --> 1x1 struct
    "A010" --> 1x1 struct
    "A011" --> 1x1 struct
    "A012" --> 1x1 struct
    "A013" --> 1x1 struct
    "A014" --> 1x1 struct
    "A015" --> 1x1 struct

d("A008")

ans = struct with fields:
    abc: 8

Answer 3

[Eric Sofen]

Is there a reason you want to use a container other than table?

Are the SerialNumbers unique? If so, you can put them in the RowNames property of the table and "query" the table using subscripting:

TestTable_valid = readtable('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1833727/TestTable_valid.xlsx')

TestTable_valid = 11×6 table
        SerialNumber        ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
    ____________________    __________________    ___________________    __________    _________    _______________

    {'A017441206210001'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A017441229220004'}         {'7.0A'}              {'8.6A'}          {'Passed'}    {'VA2C' }        {'Yes'}    
    {'A020621202230003'}         {'0A'  }              {'7.7A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621203230003'}         {'0A'  }              {'7.6A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621207230002'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621208230001'}         {'0A'  }              {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621210230006'}         {'1.2A'}              {'0.7' }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621215230003'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020621222230004'}         {'0A'  }              {'0A'  }          {'Failed'}    {'VA2CE'}        {'Yes'}    
    {'A020622108240013'}         {'7.1A'}              {'0A'  }          {'Passed'}    {'VA2CE'}        {'Yes'}    
    {'A020622109240009'}         {'7.1A'}              {'7.1A'}          {'Passed'}    {'VA2CE'}        {'Yes'}    

TestTable_valid.Properties.RowNames = TestTable_valid.SerialNumber;
TestTable_valid.SerialNumber = [];
TestTable_valid.Properties.DimensionNames(1) = "SerialNumber";

% get all data for one row
TestTable_valid("A020621208230001",:)

ans = 1×5 table
                        ACCurrentPreSyncEn    ACCurrentPostSyncEn     PassFail       Model      LCRSwept_Yes_No
                        __________________    ___________________    __________    _________    _______________

    A020621208230001          {'0A'}               {'7.1A'}          {'Failed'}    {'VA2CE'}        {'Yes'}    

% get just PassFail for one row
TestTable_valid.PassFail("A020621208230001")

ans = 1×1 cell array
    {'Failed'}

% and of course, table, unlike dictionary, means that if you need to slice
% across all SerialNumbers, you still can:
TestTable_valid.Model

ans = 11×1 cell array
    {'VA2C' }
    {'VA2C' }
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}
    {'VA2CE'}

[Walter Roberson]

Is there a reason you want to use a container other than table?

Perhaps because tables are slow?

TestTable_valid = readtable('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1833727/TestTable_valid.xlsx');
TestTable_valid.Properties.RowNames = TestTable_valid.SerialNumber;
TestTable_valid.SerialNumber = [];
TestTable_valid.Properties.DimensionNames(1) = "SerialNumber";

start = tic;
TestTable_valid("A020621208230001",:);
TestTable_valid("A017441206210001",:);
TestTable_valid("A020622109240009",:);
t1 = toc(start)

t1 = 0.0223

rownames = TestTable_valid.Properties.RowNames;
as_cell = arrayfun(@(IDX) {TestTable_valid{IDX,2:end}}, 1:height(TestTable_valid)).';
C = containers.Map(rownames, as_cell);
start = tic;
C("A020621208230001");
C("A017441206210001");
C("A020622109240009");
t2 = toc(start)

t2 = 0.0020

D = dictionary(rownames, as_cell);

start = tic;
D({'A020621208230001'});
D({'A017441206210001'});
D({'A020622109240009'});
t3 = toc(start)

t3 = 0.0020

[Jorge]

Thank you @Walter Roberson:

no....just not knowledgeable enough.....primary intent was to be able to enable the look up of an entry on a table from the key (the serial number) provided from another table of collected data. Which is the "better" way to do it?

[Eric Sofen]

@Walter Roberson, Perhaps. That's why I asked the question. You and Matt are correct that if you just need to access data by looking up the SerialNumber, the fastest thing is to use a dictionary. (And we encourage you to use dictionary rather than containers.Map. Dictionary addresses several performance and behavioral issues with containers.). Lookup performance can be particularly important when doing a fast calculation in a tight loop and dictionary probably wins there. However, as I attempted to illustrate above, if Jorge wants to do things other than look up rows, then table may still be a good option. Vectorized calculations on one variable are a lot faster and easier on the table than the dictionary of structs.

Having real-world use cases is helpful for me as I think about table performance, because there are ultimately trade-offs in performance and functionality.

[Walter Roberson]

@Eric Sofen

See by the way the timing tests I posted at https://www.mathworks.com/matlabcentral/answers/113258-is-there-any-concept-like-dictionary-or-hash-tables-in-matlab-like-in-python#comment_3254879