Convert Financial Time Series Objects fints to Timetables - MATLAB & Simulink

Convert Financial Time Series Objects `fints` to Timetables

In R2018a, financial time series (fints), and its associated methods have been replaced with a MATLAB^® timetable function. If you use fints or the associated methods, you receive a warning. You can use fts2timetable to convert a fints object to a timetable object. To help you convert from the older fints to the newer timetable functionality, use the following information.

Create Time Series

I/O Related Operations

Task	Old Functionality	New Functionality
Construct by passing in data and dates	`fints(dates,data,datanames)`	`timetable(rowTimes,var1,...,varN,'VariableNames',{'a','b',...})`
Construct by conversion of files	`ascii2fts(filename,descrow,colheadrow,skiprows)`	`T = readtable(filename,opts,Name,Value)` `TT = table2timetable(T,'RowTimes',timeVarName)`
Construct by using App with user interface	Using the Financial Time Series app with user interface	Using ImportData from HOME tab
Write files	`fts2ascii(filename,tsobj,exttext)`	`writetable(TT,filename)`
Convert to matrix	`fts2mat(tsobj)`	`S = vartype('numeric');` `TT2 = TT(:,S)` `TT2.Variables`

Index an Object

Indexing an Object

Task	Old Functionality	New Functionality
Indexing with a date	`myfts('05/11/99')`	`TT({'1999-05-11'},:)`
Indexing with a date range	`myfts ('05/11/99::05/15/99')`	`S = timerange('1999-05-11','1999-05-15');` `TT2 = TT(S,:)`
Indexing with integers for rows	`myfts.series2(1)` `myfts.series2([1, 3, 5])` `myfts.series2(16:20)`	`TT(1,{'series2'})` `TT([1, 3, 5],{'series2'})` `TT(16:20,{'series2'})`
Contents of a specific time field	`myfts.times`	`timeofday(TT.Properties.RowTimes)`
Contents for a specific field in a matrix	`fts2mat(myfts.series2)`	`TT.series2`

Transform Time Series

Assume that all variables are numeric within a timetable, or the operations can be applied on TT2:

S = vartype('numeric');

TT2 = TT(:,S)

Filter Time Series

Task	Old Functionality	New Functionality
Boxcox transformation	`newfts = boxcox(oldfts)`	`TT.Variables = boxcox(TT.Variables)`
Differencing	`diff(myfts)`	`TT{2:end,:} = diff(TT.Variables)` `TT(1,:) = []`
Indexing with integers for rows	`fillts(oldfts,fill_method)`	`fillmissing(TT,method)` (Assumes no missing dates)
Linear filtering	`filter(B,A, myfts)`	`TT.Variables = filter(b,a,TT.Variables)`
Lag or lead time series object	`lagts(myfts,lagperiod)` `leadts(myfts,leadperiod)`	`lag(TT,lagperiod)` `lag(TT,-leadperiod)` (Assumes a regularly spaced timetable)
Periodic average	`peravg(myfts)`	`retime(TT,newTimes,'mean')`
Downsample data	`resamplets(oldfts,samplestep)`	`retime(TT,newTimeStep,method)`
Smooth data	`smoothts(input)`	`smoothdata(TT)`
Moving average	`tsmovavg(tsobj,method,lag)`	`movavg(TT,type,windowSize)`

Convert Time Series

Assume that all variables are numeric within a timetable, or the operations can be applied on TT2:

S = vartype('numeric');

TT2 = TT(:,S)

Conversion Operations

Task	Old Functionality	New Functionality
Convert to specified frequency	`convertto(oldfts,newfreq)`	`retime(TT,newTimeStep,method)`
Convert to annual	`toannual(oldfts,...)`	`convert2annual(TT,…)`
Convert to daily	`todaily(oldfts,...)`	`convert2daily(TT,…)`
Convert to monthly	`tomonthly(oldfts,...)`	`convert2monthly(TT,…)`
Convert to quarterly	`toquarterly(oldfts,...)`	`convert2quarterly(TT,…)`
Convert to semiannual	`tosemi(oldfts,...)`	`convert2semiannual(TT,…)`
Convert to weekly	`toweekly(oldfts,...)`	`convert2weekly(TT,…)`

Merge Time Series

Merge Operations

Task Old Functionality New Functionality

Merge multiple time series objects

Task	Old Functionality	New Functionality
Merge multiple time series objects	`merge(fts1,fts2)`	`[TT1;TT2]` (requires variable name to be the same) `unique(TT)`
Concatenate financial time series objects horizontally	`horzcat(fts1,fts2)` or `[fts1,fts2]`	`horzcat[TT1,TT2]` (requires variable name to be the same) or `synchronize(TT1,TT2)`
Concatenate financial time series objects vertically	`vertcat(fts1,fts2)` or `[fts1;fts2]`	`vertcat[TT1;TT2]`

merge(fts1,fts2)

[TT1;TT2] (requires variable name to be the same)

unique(TT)

Concatenate financial time series objects horizontally

horzcat(fts1,fts2) or [fts1,fts2]

horzcat[TT1,TT2] (requires variable name to be the same) or

synchronize(TT1,TT2)

Concatenate financial time series objects vertically vertcat(fts1,fts2) or [fts1;fts2] vertcat[TT1;TT2]

Analyze Time Series

Due to flexibility of a timetable that can hold heterogeneous variables, a timetable does not support math operations or descriptive statistical calculations. If you would like to apply any numeric calculations on a timetable, use the following guidelines.

Assume that all variables are numeric within a timetable, or the operations can be applied on TT2:

S = vartype('numeric');

TT2 = TT(:,S)

Descriptive Statistics and Arithmetic and Math Operations

Task	Old Functionality	New Functionality
Extract out numerical data	`srs2 = myfts.series2`	`TT.Variables`
Apply some options (statistics)	For example: `min`, `max`, `mean`, `median`, `cov`, `std`, and `var`	`cov(TT.Variables)`
Apply some options (operations)	For example: `sum` and `cumsum`	`TT.Variables = cumsum(TT.Variables)`

Data Extraction

Refer to timetable documentation for data extraction methods and examples.