extract or split data from one column in table into several columns

Hi

I have a column with text data like below across multiple rows. I want to extract or split parts of this data into new columns. This data is not the same across all rows. Could anyone advise how to do this?

Ideally to get these 5 variables (separated here by /) into 5 columns

250ms,500ms / 95, 101, 106, 107 / Avg: 11_right / | bl | _HLG_MLD | (9) vs. Avg: 11_right | bl | _CG | _LOW (8) / FCL

from this data stored in one column as below

Perm t-test equal [250ms,500ms 95, 101, 106, 107]: Avg: 11_right | bl | _HLG_MLD | (9) vs. Avg: 11_right | bl | _CG | _LOW (8) | FCL

This data varies across rows but is in this same format, i.e it could be 11_right or 12_righ, etc, or could be FCL or POL, etc.

[Bob Thompson]

What format do you want the output to be in? This is fairly easy to do with split, but that's going to leave you with string outputs.

[DavidL88]

Hi Bob,

Thank you. I don't have any required format for the output.

[Bob Thompson]

Does the split command do what you need it to then?

[DavidL88]

Hi Bob

If I use InsertAfter to place / in the right positions and then use split it does. Thanks!

Answer 1

[dpb]

>> s='250ms,500ms / 95, 101, 106, 107 / Avg: 11_right / | bl | _HLG_MLD  | (9) vs. Avg: 11_right  | bl | _CG  | _LOW (8) / FCL';
>> split(s,'/')
ans =
  5×1 cell array
    {'250ms,500ms '                                                      }
    {' 95, 101, 106, 107 '                                               }
    {' Avg: 11_right '                                                   }
    {' | bl | _HLG_MLD  | (9) vs. Avg: 11_right  | bl | _CG  | _LOW (8) '}
    {' FCL'                                                              }
>> 

As Bob says, if it's only to split the string on the delimiter, it's trivial...

[DavidL88]

Hi DB

Thank you. To clarify, the '/' delimiter is not actually in the text. I put that there to illustrate the separation I wanted of this text. The data is stored as in the second example in my query (the line beginning with Perm).

[dpb]

Oh, my bad.

To have a generic routine we would then need a rule for what determines where that split belongs. This wouldn't be a fixed-width text file by any chance would it?

[DavidL88]

Hi

It’s a matlab table with 4 columns. The data I want to split is in the first column. I thought simplest approach might be to insert / where I want to split using insertafter function and then split.

[dpb]

Well, if you have a rule for what determines the string for determining the after in insertAfter that's fine. Is there a set of unique strings that can be determined to be in every string that can be used for the purpose? We only have one sample (and I'm not positive the representation in the forum is actually what it looks like in the file).

Are there no tabls, maybe, or is it actually a fixed-width file that might be simpler to just import by fixed column?