From 8c42f1fa274f8dbf53e8e89d92e16ae50a68c47e Mon Sep 17 00:00:00 2001
From: Thomas OBADIA <thomas.obadia@pasteur.fr>
Date: Thu, 17 Oct 2024 15:48:31 +0200
Subject: [PATCH] Fix some more issues introduced by dumping the databases with
categorical data as labels and a more major issue where record_id would still
be selected in subsetting some lists when it did not exist anymore.
Conversely, the censusid does not need to be calculated as part of the
get_target_list_for_observational_study() function anymore.
---
01_INVENTORY/INVENTORY_00_R_functions.R | 17 +++--------------
.../INVENTORY_01_dump_REDCap_database.R | 4 ++--
.../INVENTORY_03_list_all_inventory_houses.R | 5 ++---
...st_of_participants_for_observational_study.R | 4 ++--
4 files changed, 9 insertions(+), 21 deletions(-)
diff --git a/01_INVENTORY/INVENTORY_00_R_functions.R b/01_INVENTORY/INVENTORY_00_R_functions.R
index 9cd9c3f..6f56732 100644
--- a/01_INVENTORY/INVENTORY_00_R_functions.R
+++ b/01_INVENTORY/INVENTORY_00_R_functions.R
@@ -90,12 +90,7 @@ get_target_list_for_observational_study <- function(x, n_target, n_backup) {
# Order sampling list to appear grouped by house, for ease of use
observational_list_p_main <- bind_rows(observational_list_p_main) %>%
arrange(country, clusterid, hid, nested_hhid, nested_subjid) %>%
- mutate(censusid = paste0(country, "-",
- sprintf("%02d", clusterid), "-",
- "H", sprintf("%03d", hid), "-",
- sprintf("%02s", nested_hhid), "-",
- sprintf("%02s", nested_subjid)),
- list_name = "main")
+ mutate(list_name = "main")
# A backup list is generated after excluding people participants
# already sampled in the main list
@@ -107,19 +102,13 @@ get_target_list_for_observational_study <- function(x, n_target, n_backup) {
"n_target" = sampling_plan$nested_backup_sampling_n,
MoreArgs = list("x" = x %>%
anti_join(observational_list_p_main %>%
- select(record_id,
- country, clusterid, hid, nested_hhid, nested_subjid))),
+ select(country, clusterid, hid, nested_hhid, nested_subjid))),
SIMPLIFY = FALSE)
# Order sampling list to appear grouped by house, for ease of use
observational_list_p_backup <- bind_rows(observational_list_p_backup) %>%
arrange(country, clusterid, hid, nested_hhid, nested_subjid) %>%
- mutate(censusid = paste0(country, "-",
- sprintf("%02d", clusterid), "-",
- "H", sprintf("%03d", hid), "-",
- sprintf("%02s", nested_hhid), "-",
- sprintf("%02s", nested_subjid)),
- list_name = "backup")
+ mutate(list_name = "backup")
# Return the full list
res <- observational_list_p_main %>%
diff --git a/01_INVENTORY/INVENTORY_01_dump_REDCap_database.R b/01_INVENTORY/INVENTORY_01_dump_REDCap_database.R
index 6ba3730..ea8dbdf 100644
--- a/01_INVENTORY/INVENTORY_01_dump_REDCap_database.R
+++ b/01_INVENTORY/INVENTORY_01_dump_REDCap_database.R
@@ -110,8 +110,8 @@ if (!DATA_EXTRACT_IS_RECENT_INV) {
mutate(gps_is_manual = case_when(is.na(gps_lat) & is.na(gps_lon) & !is.na(gps_lat_manual) & !is.na(gps_lon_manual) ~ TRUE,
!is.na(gps_lat) & !is.na(gps_lon) & is.na(gps_lat_manual) & is.na(gps_lon_manual) ~ FALSE),
gps_is_valid = FALSE,
- gps_is_valid = case_when(((country == "E" & (gps_lat >= GPS_LAT_MIN_E & gps_lat <= GPS_LAT_MAX_E & gps_lon >= GPS_LON_MIN_E & gps_lon <= GPS_LON_MAX_E)) |
- (country == "M" & (gps_lat >= GPS_LAT_MIN_M & gps_lat <= GPS_LAT_MAX_M & gps_lon >= GPS_LON_MIN_M & gps_lon <= GPS_LON_MAX_M))) ~ TRUE)) %>%
+ gps_is_valid = case_when(((country %in% c("E", "Ethiopia") & (gps_lat >= GPS_LAT_MIN_E & gps_lat <= GPS_LAT_MAX_E & gps_lon >= GPS_LON_MIN_E & gps_lon <= GPS_LON_MAX_E)) |
+ (country %in% c("M", "Madagascar") & (gps_lat >= GPS_LAT_MIN_M & gps_lat <= GPS_LAT_MAX_M & gps_lon >= GPS_LON_MIN_M & gps_lon <= GPS_LON_MAX_M))) ~ TRUE)) %>%
# Add data source for convenience
mutate(data_source = RCON)
diff --git a/01_INVENTORY/INVENTORY_03_list_all_inventory_houses.R b/01_INVENTORY/INVENTORY_03_list_all_inventory_houses.R
index 7748b7e..0b0d946 100644
--- a/01_INVENTORY/INVENTORY_03_list_all_inventory_houses.R
+++ b/01_INVENTORY/INVENTORY_03_list_all_inventory_houses.R
@@ -40,11 +40,10 @@ inventory_list_h <- dat_inventory_raw %>%
# Merge in the number of individuals per house-and-household
left_join(inventory_list_p %>%
- group_by(record_id, country, clusterid, hid) %>%
+ group_by(country, clusterid, hid) %>%
summarize(n_living_at_house = n(),
.groups = "keep"),
- by = join_by("record_id" == "record_id",
- "country" == "country",
+ by = join_by("country" == "country",
"clusterid" == "clusterid",
"hid" == "hid"))
diff --git a/01_INVENTORY/INVENTORY_05_generate_list_of_participants_for_observational_study.R b/01_INVENTORY/INVENTORY_05_generate_list_of_participants_for_observational_study.R
index 06284a3..15fa9b7 100644
--- a/01_INVENTORY/INVENTORY_05_generate_list_of_participants_for_observational_study.R
+++ b/01_INVENTORY/INVENTORY_05_generate_list_of_participants_for_observational_study.R
@@ -33,8 +33,8 @@ set.seed(seed = SEED)
### GENERATE SAMPLING LIST
######################################################################
observational_list_p <- get_target_list_for_observational_study(x = inventory_list_p,
- n_target = c("E" = 220,
- "M" = 110),
+ n_target = c("Ethiopia" = 220,
+ "Madagascar" = 110),
n_backup = 50)
--
GitLab