Skip to contents

Combine a CTD sensor pair by the provider's quality flags

Source: R/sensor_pair.R
combine_sensor_pair.Rd

The rule, in precedence order:

  1. a sensor flagged 8 (questionable) or 9 (bad) is dropped;

  2. a 1 (use primary) on either sensor selects sensor 1 alone, a 2 (use secondary) selects sensor 2 alone — when both say the same thing; if they disagree (one says 1, the other 2) neither is trusted over the other and the mean is used;

  3. otherwise the mean of the two; one alone when the other is NULL; NULL when neither survives.

Usage

combine_sensor_pair(v1, v2, q1 = NA, q2 = NA)

combine_sensor_pair_sql(v1, v2, q1, q2)

Arguments

v1, v2

numeric: the two sensors' values (sensor 1 = primary).

q1, q2

their quality codes (character or numeric; NA = good).

Value

combine_sensor_pair(): a numeric vector; combine_sensor_pair_sql(): a length-one character SQL expression.

Details

Flags are compared as the source writes them — "8", "8.0", 8, NA and "" all mean what the vocabulary says — so a consumer can pass measurement_qual straight from obs. combine_sensor_pair_sql() returns the same rule as one SQL expression over four column names or expressions, for use in a DuckDB query (the transect builder, the Explorer's SQL).

Examples

combine_sensor_pair(c(10, 10, 10, 10, NA, 10), c(12, 12, 12, 12, 12, 12),
                    q1 = c(NA, "8", "1", "2", NA, "9"), q2 = c(NA, NA, NA, NA, NA, "9.0"))
#> [1] 11 12 10 12 12 NA
# 11 10 10 12 12 NA
combine_sensor_pair_sql("salinity_1_corr", "salinity_2_corr", "q1", "q2")
#> [1] "CASE WHEN COALESCE((regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') = '1' OR regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') = '1'), FALSE) AND NOT COALESCE((regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') = '2' OR regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') = '2'), FALSE) AND (CASE WHEN regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_1_corr END) IS NOT NULL THEN (CASE WHEN regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_1_corr END)\nWHEN COALESCE((regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') = '2' OR regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') = '2'), FALSE) AND NOT COALESCE((regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') = '1' OR regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') = '1'), FALSE) AND (CASE WHEN regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_2_corr END) IS NOT NULL THEN (CASE WHEN regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_2_corr END)\nWHEN (CASE WHEN regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_1_corr END) IS NOT NULL AND (CASE WHEN regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_2_corr END) IS NOT NULL THEN ((CASE WHEN regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_1_corr END) + (CASE WHEN regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_2_corr END)) / 2\nELSE COALESCE((CASE WHEN regexp_replace(CAST(q1 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_1_corr END), (CASE WHEN regexp_replace(CAST(q2 AS VARCHAR), '\\.0+$', '') IN ('8', '9') THEN NULL ELSE salinity_2_corr END)) END"