Runs of exact zeros within a cast

Finds each run of exact zeros in a series within one cast: consecutive values, in depth order, that are all 0, hold at least min_n finite values and span at least min_span_m metres. A non-zero value ends a run; so does a step in depth larger than max_gap_m between two zeros (a missing bin or two does not). It tests a different thing from check_depth_constant_series(), which asks whether the whole cast holds one value: a cast of zeros only is found by both, so a caller that withholds both applies that one first and this one to what remains. Only an exact 0 counts: a series that approaches zero (0.003, 0.001) is a measurement.

Usage

check_zero_runs(
  x,
  tbl = "ctd_measurement",
  cast_col = "ctd_cast_uuid",
  type_col = "measurement_type",
  value_col = "measurement_value",
  depth_col = "depth_m",
  types = NULL,
  min_n = 6,
  min_span_m = 50,
  max_gap_m = 5
)

Arguments

x

a DBIConnection (then tbl names the long table) or a data frame with the long-format columns below.

tbl

table name on con; ignored when x is a data frame.

cast_col, type_col, value_col, depth_col

column names: the cast key, the measurement type, the value and the depth in metres. The defaults are the CTD ingest's ctd_measurement. cast_col may name several columns, which together identify a cast (e.g. c("cruise_key", "cast_key", "cast_dir")). Group by the real cast: a key that is unique per depth scan (the CTD ingest's ctd_cast_uuid hashes the scan's time) gives one value per group, so nothing is ever judged.

types

optional character vector of measurement_types to judge; the default NULL judges every type present.

min_n

minimum values in a run (default 6).

min_span_m

minimum depth span of a run, metres (default 50).

max_gap_m

largest depth step, metres, between two zeros of one run (default 5: CTD bins are 1 m, and a step of 2-5 m is a scan or two removed).

Value

A tibble, one row per run, ordered by type, cast and depth: the cast_col column(s), measurement_type, run (1, 2, ... within the cast and type), n, depth_min_m, depth_max_m, span_m. Zero rows when none. Every value of a cast and type at a depth in [depth_min_m, depth_max_m] that is exactly 0 belongs to the run (a non-zero value there would have ended it).

Details

Which series to test is the caller's decision, and it matters. An exact 0 is a fill or a failed regression where the quantity cannot be zero (oxygen at 500 m off California beside a sensor reading 15-30 umol/kg), but it can be a real estimate clipped at zero where the quantity truly vanishes (nitrate in a depleted surface layer, chlorophyll below the chlorophyll layer); pass types accordingly.

See also

check_depth_constant_series() for a series constant over the whole cast.

Examples

d <- data.frame(cast = "a", measurement_type = "oxygen_sta_corr",
  depth_m = 0:120, measurement_value = c(seq(250, 30, length.out = 60), rep(0, 61)))
check_zero_runs(d, cast_col = "cast")
#> duckdb keeps downloaded extensions and secrets in a temporary directory:
#> ℹ /tmp/RtmpkV4Tp2/duckdb
#> This is removed when the R session ends.
#> • Extensions are re-downloaded each session.
#> • Secrets are lost.
#> ℹ Run duckdb(shared_home = TRUE) (or create ~/.duckdb) to keep them (suitable for most users).
#> ℹ Run duckdb(shared_home = FALSE) to accept the temporary directory (and silence this message).
#> ℹ See ?duckdb_storage for details and alternatives.
#> # A tibble: 1 × 7
#>   cast  measurement_type   run     n depth_min_m depth_max_m span_m
#>   <chr> <chr>            <int> <dbl>       <dbl>       <dbl>  <dbl>
#> 1 a     oxygen_sta_corr      1    61          60         120     60