How can I convert several tables into cell arrays

Hi :)

I have 6 tables in my workspace and I want to convert them all into cell arrays.

Is this possilble by a loop for example? I dont want to convert every table in an extra step

Thx :)

[dpb]

Having created N tables then wanting to turn those into a cell arrays may be indicative of there being better alternative for data storage -- perhaps to not create a table to begin with if that isn't conducive to the problem at hand or, alternatively, just not referencing the table efficiently and there isn't really a need for the cell array after all.

We haven't enough information to go on, but that's my first reaction--"why?"

[Antonio Melieni]

Hmm.. thats actually a good question. I'm a beginner.... The base dataset is a csv file. But to come back to my question, even fur the future, is it possilble to convert the tables thru a loop?

Thx :)

[Guillaume]

It's possible to convert a table to a cell array without a loop simply with table2cell. It would be a lot more complicated to do it with a loop.

The question remains why you'd want to convert a table into a cell array. Tables are a lot easier to work with than cell arrays and have a lot more functionality.

If the base dataset is a csv file of numbers, then I would understand wanting to convert the table into a matrix but not a cell array. Of course, in that case it would have been simpler to read the file directly as a matrix.

[Antonio Melieni]

Thx Guillaum for your awsner.

I m at the beginning with matlab and I have mostly no idea :)

The reason why I wanted to convert it into a cell array it that i wantet to convert the date column into datetime format.... it always showed an error...

[dpb]

As G suggested earlier, back up a step and read the original file and parse it when read it.. readtable will generally for a reasonably clean file manage to figure out the format for the variables on its own but there's the detectImportOptions function that is able to figure out a bunch more on its own as a helper...or just use textscan and write the specific format string needed. It has a %{}D format string specifically for datetime strings.

If the table imported the datetime column as a char string, then

t.Time=datetime(cellstr(t.Time));

is the basic operation you're looking for altho we would need to see what the actual form of the table is and the string format to write specifics...but you won't need to convert except transparently.

[Antonio Melieni]

EUR_USD_ASK = table2array(EURUSDCandlestick1mASK01),
EUR_USD_BID = table2array(EURUSDCandlestick1mBID01),
GBP_USD_ASK = table2array(GBPUSDCandlestick1mASK01),
GBP_USD_BID = table2array(GBPUSDCandlestick1mBID01),
USD_CHF_ASK = table2array(USDCHFCandlestick1mASK01),
USD_CHF_BID = table2array(USDCHFCandlestick1mBID01),
USD_JPY_ASK = table2array(USDJPYCandlestick1mASK01),
USD_JPY_BID = table2array(USDJPYCandlestick1mBID01);

is it right? At least i dont need to change any datetime :)

[Guillaume]

Again, it's very unclear what you are doing. Perhaps you should attach an example csv file so we understand the problem.

One thing for sure, do not create numbered variables as you have done. You should have only one variable, a cell array, a containers.Map, a structure, whatever. Something that you can easily iterate instead of having to write 8 times the same lines.

Also, up on the right of the webpage, there is a field for you to enter your version of matlab. It's particularly important to tell us which version you're using since R2019a introduced some new functions that may make reading your file a lot easier.

[dpb]

Adding to G's comment "do not create numbered variables..."

Nor sequentially-ASCII named ones with meta-date encoded in them as your LHS variables.

As far as your Q? of "right" -- the code may be syntactically correct and do what it says, but it is NOT the right solution to the problem owing to the above data storage issues...

Don't keep us in the dark guessing...show the input file (or a piece of it if it is really large) and describe briefly your objectives...

[Antonio Melieni]

First of all, thanks for your help :)

Im writing my bachelorthesis about foreign exchange rates

I have to do a timeseries analysis of minutley based data from 01.01.18-01.01.19 of four currency pairs. The datasets came as csv files.

I have to calculate several variables like the price, which for example ist the mean of (bid and ask price)/2, volatility and so on.

Attached you find a screen shot, as i couldnt upload some data....

Sorry for the stupid questions but im using matlab since 5 days and i had no lessons at all and 8 weeks to finish my thesis....

[Stephen23]

As dpb and Guillaume have already mentioned, don't stick meta-data (e.g. numbers, test-case names, parameter values, etc) into your variable names (unless you want to force yourself into writing slow, complex, buggy code that is hard to debug). Just use indexing. Indexing is simple and very efficient.

"The datasets came as csv files."

Then just follow the MATLAB documentation examples for importing multiple files:

https://www.mathworks.com/help/matlab/import_export/process-a-sequence-of-files.html

Import the data as a numeric array or a table or whatever works best, and simply store them in one cell array using indexing:

https://www.mathworks.com/help/matlab/matlab_prog/access-data-in-a-cell-array.html

"How can I convert several tables into cell arrays"

It is unlikely that this is necessary or even a good way to achieve your goal. Just access the data directly from the table (in the cell array).