Reduce portfolio by merging redundant date ranges

Transform all the date ranges together as a set to produce a new set of date ranges. Ranges separated by a gap of at least min.gapwidth days are not merged.

reduce(df, begin, end, ..., agg_cols = NULL, agg = "sum", min.gapwidth = 5)

Arguments

df: data.frame
begin: name of column df with begin dates
end: name of column in df with end dates
...: names of columns in df used to group date ranges by
agg_cols: list with columns in df to aggregate by (defaults to NULL)
agg: aggregation type (defaults to "sum")
min.gapwidth: ranges separated by a gap of at least min.gapwidth days are not merged. Defaults to 5.

Value

An object of class "reduce". The function summary is used to obtain and print a summary of the results. An object of class "reduce" is a list usually containing at least the following elements:

df: data frame with reduced time periods
begin: name of column in df with begin dates
end: name of column in df with end dates
cols: names of columns in df used to group date ranges by

Details

This function is adopted from IRanges::reduce().

Author

Martin Haringa

Examples

portfolio <- structure(list(policy_nr = c("12345", "12345", "12345", "12345",
"12345", "12345", "12345", "12345", "12345", "12345", "12345"),
productgroup = c("fire", "fire", "fire", "fire", "fire", "fire",
"fire", "fire", "fire", "fire", "fire"), product = c("contents",
"contents", "contents", "contents", "contents", "contents", "contents",
"contents", "contents", "contents", "contents"),
begin_dat = structure(c(16709,16740, 16801, 17410, 17440, 17805, 17897,
17956, 17987, 18017, 18262), class = "Date"),
end_dat = structure(c(16739, 16800, 16831, 17439, 17531, 17896, 17955,
17986, 18016, 18261, 18292), class = "Date"),
premium = c(89L, 58L, 83L, 73L, 69L, 94L, 91L, 97L, 57L, 65L, 55L)),
row.names = c(NA, -11L), class = "data.frame")

# Merge periods
pt1 <- reduce(portfolio, begin = begin_dat, end = end_dat, policy_nr,
    productgroup, product, min.gapwidth = 5)

# Aggregate per period
summary(pt1, period = "days", policy_nr, productgroup, product)
#>         date type count policy_nr productgroup  product
#> 1 2020-02-01  out     1     12345         fire contents
#> 2 2018-10-01   in     1     12345         fire contents
#> 3 2018-01-01  out     1     12345         fire contents
#> 4 2017-09-01   in     1     12345         fire contents
#> 5 2016-02-01  out     1     12345         fire contents
#> 6 2015-10-01   in     1     12345         fire contents

# Merge periods and sum premium per period
pt2 <- reduce(portfolio, begin = begin_dat, end = end_dat, policy_nr,
    productgroup, product, agg_cols = list(premium), min.gapwidth = 5)

# Create summary with aggregation per week
summary(pt2, period = "weeks", policy_nr, productgroup, product)
#>      week type count policy_nr productgroup  product
#> 1 2020W06  out     1     12345         fire contents
#> 2 2018W40   in     1     12345         fire contents
#> 3 2018W02  out     1     12345         fire contents
#> 4 2017W35   in     1     12345         fire contents
#> 5 2016W06  out     1     12345         fire contents
#> 6 2015W40   in     1     12345         fire contents