DatastreamDSWS2R

Functions and a R5 class that allows data to be downloaded and uploaded to the LSEG Datastream database via the DSWS server. LSEG Datastream is a rich database of financial information. You need to have a LSEG Datastream subscription and a username/password to use this package.

The package is installed using:

install.packages("DatastreamDSWS2R")

The development version of the package can be installed using:

require(devtools)
install_github(“CharlesCara/DatastreamDSWS2R”)

Your LSEG Datastream username and password are best stored in environment variables by adding to your .Renviron file the following lines. This approach makes DatastreamDSWS2R able to be run securely in Docker containers by passing the credentials into the container at runtime.

DatastreamUsername=YOURUSERNAME
DatastreamPassword=YOURPASSWORD

For backwards compatibility we have retained the ability to put credentials into options(). In this scenario you would add these lines to your .RProfile file:

options(Datastream.Username = "YOURUSERNAME")
options(Datastream.Password = "YOURPASSWORD")

A connection with the ‘DSWS’ API server is created by initialising the dsws object:

 mydsws <- dsws$new()

The username and password can also be passed in when creating the connection, but the environment variable approach is favoured.

 mydsws <- dsws$new(username = "YOURUSERNAME", password = "YOURPASSWORD")
 
 

Static Requests

Once set up a typical request would be for a snapshot or static request:

 mydsws <- dsws$new()
 myData <- mydsws$snapshotRequest(instrument = c("ABF","RIO","WPP"),
                               datatype = "P",
                               requestDate = "0D")

or with expressions (including E codes), where “X” is replaced by “XXXX”:

 mydsws <- dsws$new()
 myData <- mydsws$snapshotRequest(instrument = c("ABF","RIO","WPP"),
                               expression = "PCH#(XXXX,3M)",
                               requestDate = "0D")

and for a constituent list request (expressions are also supported)

mydsws <- dsws$new()
myData <- mydsws$listRequest(instrument = "LFTSE100", datatype = "P", requestDate = "0D")

We have created a set of functions that hide the reference class structure.

For example, last update can be queried either like this:

mydsws <- dsws$new()
mydsws$snapshotRequest(instrument = c("SWCNB10","UKEUSCCIR"), 
                     datatype = c("MNEM","UPDATE"), 
                     requestDate = "0D")

or like this:

mydsws <- dsws$new()
staticRequest(mydsws, c("SWCNB10","UKEUSCCIR"), c("MNEM","UPDATE"), "0D")

We can also handle composite datatypes that return multiple values, for instance datatype QTEALL, which returns details of associated stocks.

mydsws <- dsws$new()
mydsws$snapshotRequest(instrument = c("VOD", "HSBA"),
                       datatype="QTEALL",
                       requestDate = Sys.Date())

Another notable dataitem that returns multiple values is DS.USERSTATS which returns information on the user’s consumption of Datastream data.

mydsws <- dsws$new()
mydsws$snapshotRequest(instrument = "STATS",
                       datatype = "DS.USERSTATS",
                       requestDate = Sys.Date())

Timeseries requests

Timeseries request (expressions are also supported) using the timeSeriesRequest and timeSeriesListRequest methods

mydsws <- dsws$new()
xtsData <- mydsws$timeSeriesRequest(instrument = "MKS",
                       datatype = "MV",
                       startDate = "-30D",
                       endDate = "-0D",
                       frequency = "D")

xtsData <- mydsws$timeSeriesListRequest(instrument = "LFTSE100",
                       datatype = "MV",
                       startDate = "-30D",
                       endDate = "-0D",
                       frequency = "D")

Other information

The dsws interface will split large requests down into chunks small enough for the DSWS interface to process. However, the maximum chunk size varies in the DSWS documentation and is either 2000 or 50, depending on whether they have Enterprise or Individual subscriptions. The default chunkLimit is 2000, but it can be set to 50 after initialisation.

mydsws <- dsws$new()
mydsws$chunkLimit <- 50L

Alternatively this can be set as an option by adding this line to your .RProfile file.

options(Datastream.ChunkLimit = 50L)

After each data request the number of chunks and actual requests made to DSWS is saved in properties ‘numChunks’ and ‘numRequests’. This might help for data usage monitoring.

Storage of DSWS access token between sessions

The access token that is used to access DSWS can be provided via a callback function. This allows the callback function to cache the token between sessions and so reducing the number of calls on the DSWS server.

When the token expires then this function will be called again and needs to provide a refreshed token.

The callback function is used in

# Stub of function for returning token from user's cache
myTokenFunc <- function(){
    return(list(TokenValue = "abc",
                TokenExpiry = as.POSIXlt("2020-04-17 12:34")))
}

mydsws <- dsws$new(getTokenFunction = myTokenFunc)

Username and Password - ChildId

The correct username is the ‘childId’ comprising 4 characters and 3 digits. If you get a http 403 response ‘Forbidden’, then this might be because the Eikon username and password was used.

Affiliation and Warranty

This package is not provided by LSEG or an affiliate. I have written this package to help my work at Absolute Strategy Research. We are happy to make it available for others to use, but we offer or imply no warranty.

Update 1.9.1

Change to use httr::RETRY when requesting data

Update 1.8.1

Change URL to https

Update 1.7.10

Better handling of missing dates.

Update 1.7.5

Improvement in handling of Token Callback function.

Update 1.7.4

Fix for Issues #37 - charToDate error raised when list request had mixture of dates and character “NA”"

Update 1.7.2

Fix for Issues #20 - requesting token after http 403 response

Update 1.7.1

Fix for Issue #28 - callback function for getting token

Update 1.6.6

Fix for Issues #26 and #27

Update 1.6.3

Handle lower case RIC codes.

Update 1.6.2

Improve documentation and return Instrument column in snapshot requests.

Update 1.6.1

Added handling of composite datatypes which return multiple values.

Update 1.5.1

With this update we have switched from using the RCurl/rjson to using the httr/jsonlite packages for communicating with the Datastream server.

CRAN

Thank you to @mbannert for his work making the package ready to be released on CRAN.