IF cell of a dataset equal to text

Hi again.

I have a dataset with 100 rows of data with 4 columns.

I am trying to code the following and would like some help please:

For each row

If Table1.column1 = 'FOREX' (which is a string) and Table1.column4 < 0

then return Table1.column3 in Table1.column2

otherwise

If Table1.column1 = 'FOREX' (which is a string) and Table1.column4 >=0

then return Table1.column2 = (1% of Table1.column3) + Table1.column4

End

Answer 1

[dpb]

One (of many) way(s) to write it...

ix=strcmp(T.c1,'FOREX') & T.c4<0;
T(ix).c2=T(ix).c3;
ix=strcmp(T.c1,'FOREX') & T.c4>=0;
T.c2(ix)=0.01*T(ix).c3+T1(ix).c4;

Names abbreviated to protect...

[Matthew]

Brilliant thank you dpb

Answer 2

[Matthew]

The above worked fine for numeric but I'm having a similar problem when I am wanting replace a field with text in the dataset. Can you help please?

iy=or(strcmp(CtpyList.deal_type,'120'), strcmp(CtpyList.deal_type,'123'));

This works fine and creates the table iy which details 0s or 1s when it meets the above condition.

CtpyList.customer_nr(iy) = CtpyList.Customer_nr_Sec(iy);

This works but it concatenates the cell i.e. customer_nr(iy) now reads: 1 NaN

Note that NaN has been converted to string.

Any ideas on how to totally replace this string NaN with the field in the corresponding row for Customer_nr_Sec??

[dpb]

ly isn't a table, it's a logical array (vector). Need to keep class types straight or you run into the problems below.

What are the types of ...nr and ...nr_Sec, respectively? You can't mix and match types or, as you've discovered, Matlab either chokes or sometimes will try to do an implicit cast for you...

[Matthew]

Hi dpb,

Yes sorry - iy is a logical array.

customer_nr and Customer_nr_Sec are column headers in datasets and the data in these cells are both string (even the NaN which is currently present in customer_nr due to a num2str performed in a previous script).

[dpb]

Probably your problem is that they're character strings, maybe? I can't see your data itself but what does

CtpyList.customer_nr(iy,:) = CtpyList.Customer_nr_Sec(iy,:);

do? Does that solve the problem?

If not, post back with the results of

whos CtpyList.customer_nr(iy)
whos CtpyList.Customer_nr_Sec(iy)

for some location in iy -- say the first T would be ok.

[Matthew]

dpb - you beauty!

This first line of code you sent me does the trick - thank you again!

[dpb]

Good; thought it might.

You do understand why, then???

[Matthew]

Yes