Reading a string & sorting it into usable variables

Hi!

I am working on a function that reads a .lpt file & then completes some computations using the values included.

This is what the test file structure looks like:

"

Layup: [45/-45/0/90/90/0/-45/45]

Ply thickness: [0.18/0.18/0.18/0.18/0.18/0.18/0.18/0.18]

Elastic properties: "IM7", 162000, 8340, 2070, 0.34

"

I can read the entire file in as a string. What is the best method to take that string & remove the vectors & properties from it? Essentially, I want to delete the "Layup:" & save the values in [ ] as a vector without the /. The lengths of the vectors can change too.

Thank you!

Answer 1

[Steven Lord]

Let's take a sample string.

S = "Layup: [45/-45/0/90/90/0/-45/45]";

Get the text after the leading [

S2 = extractAfter(S, "[")

S2 = "45/-45/0/90/90/0/-45/45]"

and before the trailing ].

S3 = extractBefore(S2, "]")

S3 = "45/-45/0/90/90/0/-45/45"

Now we could split S3 into pieces using the slash character as the delimiter then convert to double.

S4 = split(S3, "/")

S4 = 8×1 string array
    "45"
    "-45"
    "0"
    "90"
    "90"
    "0"
    "-45"
    "45"

S4 = double(S4).'

S4 = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Alternately, replace the / with space and use sscanf.

S5 = replace(S3, "/", " ")

S5 = "45 -45 0 90 90 0 -45 45"

S5 = sscanf(S5, "%d").'

S5 = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

It's not clear how you want to handle the 'Elastic properties' line: is "IM7" data to be saved or 'header' / 'description' to be removed? I'm guessing data, but you can't have both text and numbers in a numeric array.

Answer 2

[Stephen23]

Create demo file:

writelines({'Layup: [45/-45/0/90/90/0/-45/45]','Ply thickness: [0.18/0.18/0.18/0.18/0.18/0.18/0.18/0.18]','Elastic properties: "IM7", 162000, 8340, 2070, 0.34'},'./mydata.ipt')

Check file content:

type ./mydata.ipt

Layup: [45/-45/0/90/90/0/-45/45]
Ply thickness: [0.18/0.18/0.18/0.18/0.18/0.18/0.18/0.18]
Elastic properties: "IM7", 162000, 8340, 2070, 0.34

Import file data as a table:

tbl = readtable('./mydata.ipt', 'FileType','text','ReadVariableNames',false, 'ReadRowNames',true, 'Delimiter',':');
tbl.Properties.VariableNames = {'Value'}

tbl = 3×1 table
                                              Value                    
                          _____________________________________________

    Layup                 {'[45/-45/0/90/90/0/-45/45]'                }
    Ply thickness         {'[0.18/0.18/0.18/0.18/0.18/0.18/0.18/0.18]'}
    Elastic properties    {'IM7, 162000, 8340, 2070, 0.34'            }

Option 1: one non-scalar structure

nss = struct('Name',tbl.Properties.RowNames);
for k = 1:height(tbl)
    [nss(k).Text,nss(k).Value] = parseval(tbl.Value{k});
end
display(nss)

nss = 3×1 struct array with fields:
    Name
    Text
    Value

This you can access using a combination of structure indexing, e.g.:

nss(3).Name

ans = 'Elastic properties'

nss(3).Text

ans = 'IM7'

nss(1).Value

ans = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Option 2: nested scalar structures

ss = struct();
for k = 1:height(tbl)
    f = matlab.lang.makeValidName(tbl.Properties.RowNames{k});
    [ss.(f).Text,ss.(f).Value] = parseval(tbl.Value{k});
end
display(ss)

ss = struct with fields:
                Layup: [1×1 struct]
         PlyThickness: [1×1 struct]
    ElasticProperties: [1×1 struct]

This you can access using dot indexing into the nested structures, e.g.:

ss.ElasticProperties.Text

ans = 'IM7'

ss.Layup.Value

ans = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Option 3: one table

for k = 1:height(tbl)
    [tbl.Text{k},tbl.Value{k}] = parseval(tbl.Value{k});
end
display(tbl)

tbl = 3×2 table
                                                     Value                                  Text   
                          ___________________________________________________________    __________

    Layup                 {[                                45 -45 0 90 90 0 -45 45]}    {0×0 char}
    Ply thickness         {[0.1800 0.1800 0.1800 0.1800 0.1800 0.1800 0.1800 0.1800]}    {0×0 char}
    Elastic properties    {[                                162000 8340 2070 0.3400]}    {'IM7'   }

You could also transpose the table so that you can access the name using dot indexing:

tbl = rows2vars(tbl,'VariableNamingRule','modify');
tbl.Properties.RowNames = tbl.OriginalVariableNames; % this should be the default of ROWS2VARS
tbl = removevars(tbl,'OriginalVariableNames') % together with this.

tbl = 2×3 table
                        Layup                                      PlyThickness                                 ElasticProperties     
             ___________________________    ___________________________________________________________    ___________________________

    Value    {[45 -45 0 90 90 0 -45 45]}    {[0.1800 0.1800 0.1800 0.1800 0.1800 0.1800 0.1800 0.1800]}    {[162000 8340 2070 0.3400]}
    Text     {0×0 char                 }    {0×0 char                                                 }    {'IM7'                    }

Access the data using table indexing, e.g.:

tbl.ElasticProperties{2}

ans = 'IM7'

tbl.Layup{1}

ans = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

tbl{"Value","Layup"}{1}

ans = 1×8

45 -45 0 90 90 0 -45 45

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Function which you can easily modify as required, to suit different data formats:

function [str,num] = parseval(txt)
[one,~,~,idx] = sscanf(txt,'[%f');
str = '';
if idx>1
    num = [one,sscanf(txt(idx:end),'/%f',[1,Inf])];
else
    [str,~,~,idx] = sscanf(txt,'%[^,]');
    num = sscanf(txt(idx:end),',%f',[1,Inf]);
end
end