Order of BDH output #139
Comments
|
This was previously discussed in another issue, and dismissed. Please search old issue before (effectively re-)reporting the same. |
|
Bloomberg provides a sequenceNumber value in each securityData element of the response. This represents the position of that returned security in the original array of requested securities. This could be used to construct a resultant data frame following the requested order. |
|
Interesting. PRs welcome. |
R> Rblpapi::bdh(securities = c("SPX Index", "CCMP Index"), fields = "PX_LAST", start.date = Sys.Date() - 10, end.date = Sys.Date() - 10, verbose=TRUE)
HistoricalDataResponse = {
securityData = {
security = "SPX Index"
eidData[] = {
}
sequenceNumber = 0
fieldExceptions[] = {
}
fieldData[] = {
fieldData = {
date = 2016-02-19
PX_LAST = 1917.780000
}
}
}
}
HistoricalDataResponse = {
securityData = {
security = "CCMP Index"
eidData[] = {
}
sequenceNumber = 1
fieldExceptions[] = {
}
fieldData[] = {
fieldData = {
date = 2016-02-19
PX_LAST = 4504.429000
}
}
}
}
$`SPX Index`
date PX_LAST
1 2016-02-19 1917.78
$`CCMP Index`
date PX_LAST
1 2016-02-19 4504.43
R> |
|
Seems messy to take care of auxiliary data fields but if somebody has an itch to scratch ... |
|
bds needs to be refactored eventually to use two passes, one to capture the fields (as bbg does not provide a schema or fieldInfo for bds requests), and one to populate the result using the full rank fieldspace (I hope that makes sense, I can elaborate if I explained that poorly). So, when that refactoring is done, we can capture the results into a vector by seqNumber which should map to the order of the user request. It's similar to what was done here: and a check to make sure we match the user request (in case of errors in the security vector, I didn't know how bbg would handle this) Anyway, this will be an easy fix in terms of the refactoring we need to do. |
I am using
bdhto pull the closing prices of equity indexes. I have noticed that the output order does not correspond to the security order input in the function. For example:Interestingly the output is not the same order as the input:
(manually edited out BB price data)
I tried looking into the R and C++ code but I cannot seem to find where it goes wrong. It looks like it happens in the C++ implementation of
bdh_Impl. Perhaps Rcpp just casts it as a list and then R reorganises?The text was updated successfully, but these errors were encountered: