Quick Start

This chapter goes over installation instructions and one example call using the EPFR API.

API Install

To install the EPFR API library epfr, you can find the code to do this below.

# to install the EPFR API package run this line
install.packages ("epfr", repos = "https://downloads.epfrglobal.com/sdks/R/", type = "source")   
library(epfr)

API Login

Please use your EPFR credentials to login to the epfr API.
If you need help with your credentials you can contact the EPFR support team

username <- "EPFR username"
password <- "EPFR password"

epfr.login(username, password)

## [1] "Environment set to: PROD"

## [1] "Previous session closed."

## [1] "Login successful."

API Call

The process of pulling data from the EPFR API follows 3 steps,
1. Create a report using the function epfr.get.reportid(). Parameter inputs can be tailored for desired output.
2. Wait for the API to complete the report.
3. Retrieve data from the report using the function epfr.get.reportoutput().

Below is the function which which checks if a report has been completed and if it’s data is ready to be retrieved.

wait_for_completion <- function(report_id, dataset, inital_time_offset=5, time_offset=10) {
  
  time_taken <- 0
  
  print(paste0("Inital wait for report to complete, delay ", inital_time_offset))
  Sys.sleep(inital_time_offset)
  time_taken <- time_taken + inital_time_offset
  
  st <- epfr.get.reportstatus(dataset, report_id)$data$status # CHECK STATUS
  
  while (st != "Completed") {
    st <- epfr.get.reportstatus(dataset, report_id)$data$status
    
    if (st=="Completed") { 
      break # REPORT COMPLETE
    } else if (st %in% c("Errored","Failed")) {
      stop("EPFR job failed") # REPORT FAILED
    }
    
    Sys.sleep(time_offset)
    time_taken <- time_taken + time_offset
    print(paste0("Report ", st, " time taken: ", time_taken))
  } 
  
  print(paste0("Report complete: ", time_taken, " seconds taken"))
}

Example API call

Below is a call to get the total net assets for all equity ETFs in the energy sector from the fund flows database between January 2025 and May 2025.
More information about the available filters, categories, and asset classes in the API Dictionaries section of this notebook.

library(dplyr)

dataset <- "FF"

report <- epfr.get.reportid(dataset = dataset, 
                            from_date = "01-01-2025",
                            to_date = "05-01-2025",
                            asset_classes = list("FF100000"), # ALL EQUITY
                            frequency = "Daily", 
                            level = 0, # AGGREGATE
                            categories = list(1), # TOTAL NET ASSETS 
                            average_type = NULL, 
                            allocation_ids = NULL, 
                            assets = NULL, 
                            universal_filters = list("U100001"), # ETF
                            equity_filters = list("E100006"), # ENERGY
                            bond_filters = NULL, 
                            alternative_filters = NULL)  

report_id <- report$data$reportId  
wait_for_completion(report_id, dataset)
result <- epfr.get.reportoutput(dataset, report_id)  

## [1] "Inital wait for report to complete, delay 5"

## [1] "Report complete: 5 seconds taken"

Output from the example API call.

print(result %>% top_n(10))  

Date	Asset Class	Filters	Total Net Assets
2/20/2025	All Equity-FF-Equity	ETF Only, Energy	104413.3
2/19/2025	All Equity-FF-Equity	ETF Only, Energy	103783.8
1/24/2025	All Equity-FF-Equity	ETF Only, Energy	103957.8
1/23/2025	All Equity-FF-Equity	ETF Only, Energy	104698.3
1/22/2025	All Equity-FF-Equity	ETF Only, Energy	104484.4
1/21/2025	All Equity-FF-Equity	ETF Only, Energy	105826.6
1/20/2025	All Equity-FF-Equity	ETF Only, Energy	106284.1
1/17/2025	All Equity-FF-Equity	ETF Only, Energy	106260.8
1/16/2025	All Equity-FF-Equity	ETF Only, Energy	105720.3
1/15/2025	All Equity-FF-Equity	ETF Only, Energy	104548.1

API Date Format

The API accepts dates in the form “mm-dd-YYYY” whereas flow dates are written “YYYYmmdd”.
and The API outputs dates in the form “m/d/YYYY” which does not sort chronologically.
Below are 2 helper functions to convert flow dates to the api input dates and api output dates to flow dates.

flow_to_api_date <- function(flow_date) {
  y <- substr(flow_date, 1, 4) # INPUT FORMAT "YYYYMMDD"
  m <- substr(flow_date, 5, 6)
  d <- substr(flow_date, 7, 8)
  return(paste0(m, "-", d, "-", y)) #OUTPUT FORMAT "MM-DD-YYYY"
}

chrdate_to_flowdate <- function(date) {
  split <- strsplit(date, split="/")[[1]] # INPUT FORMAT "M/D/YYYY"
  y <- split[3]
  m <- ifelse(nchar(split[1]) == 1, paste0("0", split[1]), split[1])
  d <- ifelse(nchar(split[2]) == 1, paste0("0", split[2]), split[2])
  return(paste0(y, m, d)) # OUTPUT FORMATT "YYYYMMDD"
}
chrdate_to_flowdate_vec <- Vectorize(chrdate_to_flowdate)

Supplemental Libraries

We use the below libraries from tidyverse and EPFR. Please see EPFR.r Library for more information on installing the EPFR.r library.

library(EPFR.r)
library(tidyr)
library(dplyr)

---