The following examples use the following table_string helper to succintly define tables:
# Convert a string into a data.frame
table_string <- function (text, ...) read.table(
text = text,
blank.lines.skip = TRUE,
header = TRUE,
stringsAsFactors = FALSE,
...)Firstly, connect to a database and set up some areas/divisions:
mdb <- mfdb(tempfile(fileext = '.duckdb'))
mfdb_import_area(mdb, table_string('
name division size
45G01 divA 10
45G02 divA 200
45G03 divB 400
'))Populations are arbitrary groupings for defining logical stocks that canโt be derived from other MFDB data.
As with other metadata, we have to import valid values before using:
mfdb_import_population_taxonomy(mdb, table_string('
name description t_group
ns "Northern Shrimp" ns
ns_s "Northern Shrimp in Skjalfandi" ns
ns_a "Northern Shrimp in Arnarfjordur" ns
ns_i "Northern Shrimp in Isafjardardjup" ns
as "Aesop Shrimp" as
as_s "Aesop Shrimp in Skjalfandi" as
'))Notice that we have used the t_group column to define groupings of within our population groups. This means that the ns group will include all samples from ns, ns_s`,ns_a,ns_i``.
Now we can import data that uses these groupings:
mfdb_import_survey(mdb, data_source = "x",
table_string("
year month areacell species population length count
2019 1 45G01 PRA ns_s 10 285
2019 1 45G01 PRA ns_s 20 273
2019 1 45G01 PRA ns_a 10 299
2019 1 45G01 PRA ns_a 20 252
2019 1 45G01 PRA ns_i 10 193
2019 1 45G01 PRA ns_i 20 322
"))We can now use the mfdb_sample_* functions to select this data back out again. We can group and filter by any of the tow attributes. We can query for individual fjords as well as the whole group:
agg_data <- mfdb_sample_count(mdb, c('population', 'length'), list(
population = mfdb_group(ns_s = 'ns_s', ns = 'ns'),
length = mfdb_unaggregated()))
agg_data## $`0.0.0.0.0`
## year step area population length number
## 1 all all all ns 10 777
## 2 all all all ns 20 847
## 3 all all all ns_s 10 285
## 4 all all all ns_s 20 273mfdb_disconnect(mdb)