Branch over sf objects #713

lindsayplatt · 2021-11-29T23:50:03Z

lindsayplatt
Nov 29, 2021

I would like to dynamically branch over an sf object to use each individual row's geometry as an input to a web query. I don't understand why there is an error when doing this? In my mind, it should operate very similarly to mapping over a data.frame. I want to branch over the geometries so that I only re-download data if the polygon updates.

Here is an example _targets.R file for this use case.

library(targets)
options(tidyverse.quiet = TRUE)
tar_option_set(packages = c(
  "tidyverse",
  "sf",
  "maps"
))

 list(

  tar_target(
    states_sf,
    st_as_sf(maps::map('state', c('ohio', 'virginia', 'missouri'), plot=F, fill=T))
  ),

  # I can't seem to map over sf objects. Why not?
  tar_target(
    state_map_error, {
      png(sprintf("map_%s.png", states_sf), width = 500, height = 700)
      plot(states_sf)
      dev.off()
    },
    pattern = map(states_sf)
  ),

  # Works if I map over a data.frame without the geometries
  tar_target(
    states, st_drop_geometry(states_sf),
  ),
  tar_target(
    state_map_success, {
      png(sprintf("map_%s.png", states$ID), width = 500, height = 700)
      plot(filter(states_sf, ID == states$ID))
      dev.off()
    },
    pattern = map(states)
  )

)

When running tar_make(state_map_error), I get the following error:

Error : Input must be a vector, not a <sfc_MULTIPOLYGON/sfc> object.

wlandau · 2021-11-30T02:10:58Z

wlandau
Nov 30, 2021
Maintainer

That's odd, your first pipeline works for me and produces reasonable-looking PNG maps of states. It looks like it may come from vctrs, but I doubt vctrs itself is the problem because targets requires vctrs >= 0.3.8 (current CRAN). Maybe something changed in sf? For me, class(tar_read(states_sf)) is c("sf", "data.frame"), and there is no "sfc_MULTIPOLYGON" or "sfc" class. Those classes would have to inherit from data.frame or support their own [ operators.

3 replies

wlandau Nov 30, 2021
Maintainer

For what it's worth, I have sf 1.0.2 and maps 3.4.0.

lindsayplatt Nov 30, 2021
Author

Interesting ... I have the same versions (except for maps) and the most recent targets. I am just going to try to update all my packages for now and see what happens.

sapply(c("sf", "maps", "vctrs", "targets"), packageDescription, fields = "Version")

     sf    maps   vctrs targets 
"1.0-2" "3.3.0" "0.3.8" "0.8.1"

lindsayplatt Nov 30, 2021
Author

I tried updating the maps package so that I had your exact versions of these packages, but the state_map_error target still didn't build. I went ahead and updated all packages that may have a minor version change (there were 61), so it could have been any one of these:

c("arrow", "backports", "brio", "broom", "cli", "cpp11", "crayon", "credentials", "crosstalk",
  "crul", "data.table", "desc", "devtools", "diffobj", "digest", "generics", "gert", "glmnet",
  "glue", "hms", "igraph", "knitr", "labelled", "lifecycle", "lubridate", "lwgeom", "memoise", 
  "mime", "mvtnorm", "ncdf4", "odbc", "optparse", "pillar", "pkgload", "raster", "rcmdcheck",
  "RcppArmadillo", "readr", "remotes", "rgeos", "rlang", "rvest", "s2", "satellite", "servr",
  "sessioninfo", "sf", "sp", "stringi", "systemfonts", "testthat", "tibble", "tidyr", "tinytex",
  "tzdb", "usethis", "uuid", "viridis", "vroom", "xfun")

Perhaps tidyr or tibble? Well, it is working now so I guess, once again, the moral of the story is to update your packages frequently.

hansvancalster · 2022-03-16T17:24:04Z

hansvancalster
Mar 16, 2022

The example provided by the OP works for me too, but I have a different pipeline where I run into the same error:

Error in tar_throw_run():
! callr subprocess failed: could not load dependencies of target punten_34e41e49. Input must be a vector, not a <sfc_POLYGON/sfc> object.

The strange thing is that the error does not occur the first time tar_make() is called. But the second time, without changes, one of the targets starts to rebuild and fails with the above error. The target that fails uses dynamic branching and maps over two sf objects:

class(tar_read(paden))
[1] "sf"         "data.frame"
> class(tar_read(perimeters_data))
[1] "sf"         "data.frame"

I could not even reproduce the error in debug mode. Using tar_workspace(), the failing target runs just fine. I also recently updated all my packages.
I suspect that this will not be sufficient information to be able to help me out, but I haven't been able to figure out how to make a minimal reprex where this error happens.

6 replies

hansvancalster Mar 16, 2022

I'll give it a try tomorrow and report back.

hansvancalster Mar 17, 2022

Unfortunately, that didn't solve the problem. I made sure every input and output of a target was of class data.frame (where it used to be of class sf and data.frame).
The same error pops up. I believe this is because the geometry column in the plain data.frame object still is of class sfc_POLYGON:

> class(perimeters_data)
[1] "data.frame"
> perimeters_data$geom
Geometry set for 1 feature 
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: 176571.9 ymin: 154102.1 xmax: 243089.1 ymax: 173906.8
Projected CRS: Belge 1972 / Belgian Lambert 72
POLYGON ((176660.8 166141.8, 176624.3 166158.8,...
> class(perimeters_data$geom)
[1] "sfc_POLYGON" "sfc"

For my use case, I cannot just remove the geom column because the functions do spatial stuff.

hansvancalster Mar 17, 2022

I think the issue is that some subprocess has no access to the sf package when I rerun tar_make() the second time. I can "reproduce" the error when I restart my R session and then:

> library(targets)
Warning message:
package ‘targets’ was built under R version 4.1.3 
> library(vctrs)
> vec_slice(tar_read(perimeters_data), 1)
Error in `stop_vctrs()`:
! Input must be a vector, not a <sfc_POLYGON/sfc> object.
Run `rlang::last_error()` to see where the error occurred.

When I restart again and load sf, the error is gone:

> library(targets)
Warning message:
package ‘targets’ was built under R version 4.1.3 
> library(vctrs)
> library(sf)
Linking to GEOS 3.9.1, GDAL 3.2.1, PROJ 7.2.1; sf_use_s2() is TRUE
Warning message:
package ‘sf’ was built under R version 4.1.3 
> vec_slice(tar_read(perimeters_data), 1)
                   Naam                           geom
1 Oostelijke leemstreek POLYGON ((176660.8 166141.8...

My tar_option_set has sf loaded and looks like this:

tar_option_set(packages = c("tidyverse", "sf", "nngeo", "readxl", "httr",
                            "osmextract", "terra", "exactextractr", "arrow"),
               format = "qs",
               memory = "transient",
               garbage_collection = TRUE,
               workspace_on_error = TRUE)

wlandau Mar 17, 2022
Maintainer

The order packages are loaded can affect this. Maybe try putting "sf" first in packages, then if that doesn't work, try putting it last.

hansvancalster Mar 18, 2022

I've placed "sf" last so it's functions don't get masked by other packages, but that did not cause the error (I've seen it pop up afterwards).
The unexpected rebuild on the other hand, was totally my bad. I had forgotten to remove a debug = ... in tar_option_set().

hansvancalster · 2022-03-18T15:23:07Z

hansvancalster
Mar 18, 2022

The following reprex shows two issues with branching (iteration = "vector") over objects of class "sf" - "data.frame":

tar_read() of the aggregated target (by_row in the reprex) fails, unless library(sf) is loaded
making a change to the command of the target that branches over the object of class "sf" - "data.frame" will trigger a rebuilt of the target that fails when the libraries are loaded via tar_option_set(packages = ...). This does not happen when the libraries are loaded via library() calls in the target script.

library(targets)
#> Warning: package 'targets' was built under R version 4.1.3
tar_script({
    tar_option_set(packages = c("dplyr", "sf"))
    example_sf_file <- function() {
        system.file("shape/nc.shp", package = "sf")
    }
    read_shape <- function(shapefile) {
        nc <- st_read(shapefile) %>% head()
        nc
    }
    list(tarchetypes::tar_file(nc_file, example_sf_file()), tar_target(nc, 
        read_shape(shapefile = nc_file), format = "qs"), tar_target(by_row, 
        nc, pattern = map(nc), format = "qs"))
})
tar_make()
#> * start target nc_file
#> * built target nc_file
#> * start target nc
#> Reading layer `nc' from data source `C:\R\library\sf\shape\nc.shp' using driver `ESRI Shapefile'
#> Simple feature collection with 100 features and 14 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS:  NAD27
#> * built target nc
#> * start branch by_row_c8040484
#> * built branch by_row_c8040484
#> * start branch by_row_bd44f74a
#> * built branch by_row_bd44f74a
#> * start branch by_row_58077b70
#> * built branch by_row_58077b70
#> * start branch by_row_9ff1fc0f
#> * built branch by_row_9ff1fc0f
#> * start branch by_row_c28ade2e
#> * built branch by_row_c28ade2e
#> * start branch by_row_bfe79c7e
#> * built branch by_row_bfe79c7e
#> * built pattern by_row
#> * end pipeline: 0.55 seconds
#> Warning messages:
#> 1: package 'targets' was built under R version 4.1.3 
#> 2: package 'dplyr' was built under R version 4.1.3 
#> 3: package 'sf' was built under R version 4.1.3 
#> 4: 1 targets produced warnings. Run tar_meta(fields = warnings) for the messages. 
tar_read(nc)
#>    AREA PERIMETER CNTY_ CNTY_ID        NAME  FIPS FIPSNO CRESS_ID BIR74 SID74
#> 1 0.114     1.442  1825    1825        Ashe 37009  37009        5  1091     1
#> 2 0.061     1.231  1827    1827   Alleghany 37005  37005        3   487     0
#> 3 0.143     1.630  1828    1828       Surry 37171  37171       86  3188     5
#> 4 0.070     2.968  1831    1831   Currituck 37053  37053       27   508     1
#> 5 0.153     2.206  1832    1832 Northampton 37131  37131       66  1421     9
#> 6 0.097     1.670  1833    1833    Hertford 37091  37091       46  1452     7
#>   NWBIR74 BIR79 SID79 NWBIR79
#> 1      10  1364     0      19
#> 2      10   542     3      12
#> 3     208  3616     6     260
#> 4     123   830     2     145
#> 5    1066  1606     3    1197
#> 6     954  1838     5    1237
#>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       geometry
#> 1                                                                                                                                                                                                                                        -81.47276, -81.54084, -81.56198, -81.63306, -81.74107, -81.69828, -81.70280, -81.67000, -81.34530, -81.34754, -81.32478, -81.31332, -81.26624, -81.26284, -81.24069, -81.23989, -81.26424, -81.32899, -81.36137, -81.36569, -81.35413, -81.36745, -81.40639, -81.41233, -81.43104, -81.45289, -81.47276, 36.23436, 36.27251, 36.27359, 36.34069, 36.39178, 36.47178, 36.51934, 36.58965, 36.57286, 36.53791, 36.51368, 36.48070, 36.43721, 36.40504, 36.37942, 36.36536, 36.35241, 36.36350, 36.35316, 36.33905, 36.29972, 36.27870, 36.28505, 36.26729, 36.26072, 36.23959, 36.23436
#> 2                                                                                                                                                                                                                                                             -81.23989, -81.24069, -81.26284, -81.26624, -81.31332, -81.32478, -81.34754, -81.34530, -80.90344, -80.93355, -80.96577, -80.94967, -80.95639, -80.97795, -80.98284, -81.00278, -81.02464, -81.04280, -81.08425, -81.09856, -81.11331, -81.12938, -81.13840, -81.15337, -81.17667, -81.23989, 36.36536, 36.37942, 36.40504, 36.43721, 36.48070, 36.51368, 36.53791, 36.57286, 36.56521, 36.49831, 36.46722, 36.41473, 36.40380, 36.39138, 36.37183, 36.36668, 36.37783, 36.41034, 36.42992, 36.43115, 36.42285, 36.42633, 36.41763, 36.42474, 36.41544, 36.36536
#> 3                                                                                                                                                                                                                   -80.45634, -80.47639, -80.53688, -80.54501, -80.55415, -80.59060, -80.62431, -80.66744, -80.69666, -80.72404, -80.73437, -80.75256, -80.76630, -80.78270, -80.87438, -80.87086, -80.88892, -80.92457, -80.95639, -80.94967, -80.96577, -80.93355, -80.90344, -80.83816, -80.61105, -80.43531, -80.45301, -80.45634, 36.24256, 36.25473, 36.25674, 36.27666, 36.27843, 36.26828, 36.27310, 36.24610, 36.25909, 36.25847, 36.26476, 36.25830, 36.26184, 36.24858, 36.23388, 36.32462, 36.35443, 36.37275, 36.40380, 36.41473, 36.46722, 36.49831, 36.56521, 36.56344, 36.55730, 36.55104, 36.25709, 36.24256
#> 4 -76.00897, -76.01735, -76.03288, -76.04395, -76.09509, -76.16093, -76.15815, -76.16829, -76.33025, -76.12740, -76.04596, -76.03321, -76.09106, -75.97607, -75.96976, -76.00161, -75.95126, -75.92812, -75.92459, -75.80006, -75.79885, -75.85516, -75.91376, -75.95751, -75.94193, -76.00897, 36.31960, 36.33773, 36.33598, 36.35359, 36.34892, 36.39190, 36.41269, 36.42709, 36.55606, 36.55716, 36.55695, 36.51437, 36.50357, 36.43621, 36.41512, 36.41891, 36.36547, 36.42324, 36.35095, 36.11282, 36.07282, 36.10567, 36.24480, 36.25945, 36.29434, 36.31960, -76.02717, -75.99866, -75.91192, -75.92480, -75.97728, -75.97629, -76.02717, 36.55672, 36.55665, 36.54253, 36.47398, 36.47802, 36.51793, 36.55672, -75.90199, -75.87817, -75.77316, -75.78317, -75.90199, 36.55620, 36.55587, 36.22926, 36.22519, 36.55620
#> 5                                                                                     -77.21767, -77.23461, -77.29861, -77.29351, -77.30948, -77.33499, -77.38903, -77.37862, -77.41347, -77.41885, -77.45412, -77.53808, -77.55743, -77.57195, -77.58008, -77.55968, -77.56042, -77.63597, -77.65099, -77.69888, -77.74940, -77.90121, -77.89886, -77.76393, -77.32005, -77.17735, -77.15442, -77.09213, -77.07531, -77.08327, -77.12733, -77.13932, -77.14196, -77.21767, 36.24098, 36.21460, 36.21153, 36.16286, 36.16277, 36.18285, 36.20310, 36.24009, 36.25582, 36.28180, 36.31976, 36.30246, 36.30421, 36.31450, 36.32827, 36.37595, 36.40636, 36.44054, 36.47259, 36.47896, 36.47357, 36.50989, 36.55294, 36.55344, 36.55392, 36.55629, 36.52625, 36.50752, 36.48352, 36.47010, 36.47071, 36.45648, 36.41706, 36.24098
#> 6                                                                                                                                                                                                                                                                                                                                                 -76.74506, -76.98069, -76.99475, -77.13007, -77.21767, -77.14196, -77.13932, -77.12733, -77.08327, -77.07531, -77.09213, -77.15442, -77.17735, -76.92413, -76.90842, -76.94577, -76.95367, -76.94351, -76.92408, -76.74135, -76.70750, -76.74506, 36.23392, 36.23024, 36.23558, 36.23346, 36.24098, 36.41706, 36.45648, 36.47071, 36.47010, 36.48352, 36.50752, 36.52625, 36.55629, 36.55415, 36.50428, 36.45896, 36.41923, 36.40173, 36.39244, 36.31517, 36.26613, 36.23392
class(tar_read(nc))
#> [1] "sf"         "data.frame"
tar_read(by_row)
#> Error in `stop_vctrs()`:
#> ! Input must be a vector, not a <sfc_MULTIPOLYGON/sfc> object.
tar_script({
    tar_option_set(packages = c("dplyr", "sf"))
    example_sf_file <- function() {
        system.file("shape/nc.shp", package = "sf")
    }
    read_shape <- function(shapefile) {
        nc <- st_read(shapefile) %>% head()
        nc
    }
    list(tarchetypes::tar_file(nc_file, example_sf_file()), tar_target(nc, 
        read_shape(shapefile = nc_file), format = "qs"), tar_target(by_row, 
        nc %>% select(AREA), pattern = map(nc), format = "qs"))
})
tar_make()
#> v skip target nc_file
#> v skip target nc
#> * start branch by_row_c8040484
#> x error branch by_row_c8040484
#> * end pipeline: 0.5 seconds
#> Error in tar_throw_run(target$metrics$error) : 
#>   could not load dependencies of target by_row_c8040484. Input must be a vector, not a <sfc_MULTIPOLYGON/sfc> object.
#> In addition: Warning messages:
#> 1: package 'targets' was built under R version 4.1.3 
#> 2: package 'dplyr' was built under R version 4.1.3 
#> 3: package 'sf' was built under R version 4.1.3 
#> 4: 1 targets produced warnings. Run tar_meta(fields = warnings) for the messages. 
#> Error in `tar_throw_run()`:
#> ! callr subprocess failed: could not load dependencies of target by_row_c8040484. Input must be a vector, not a <sfc_MULTIPOLYGON/sfc> object.
#> Visit https://books.ropensci.org/targets/debugging.html for debugging advice.
tar_script({
    library(dplyr)
    library(sf)
    example_sf_file <- function() {
        system.file("shape/nc.shp", package = "sf")
    }
    read_shape <- function(shapefile) {
        nc <- st_read(shapefile) %>% head()
        nc
    }
    list(tarchetypes::tar_file(nc_file, example_sf_file()), tar_target(nc, 
        read_shape(shapefile = nc_file), format = "qs"), tar_target(by_row, 
        nc %>% select(AREA), pattern = map(nc), format = "qs"))
})
tar_make()
#> 
#> Attaching package: 'dplyr'
#> 
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> 
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
#> 
#> Linking to GEOS 3.9.1, GDAL 3.2.1, PROJ 7.2.1; sf_use_s2() is TRUE
#> v skip target nc_file
#> v skip target nc
#> * start branch by_row_c8040484
#> * built branch by_row_c8040484
#> * start branch by_row_bd44f74a
#> * built branch by_row_bd44f74a
#> * start branch by_row_58077b70
#> * built branch by_row_58077b70
#> * start branch by_row_9ff1fc0f
#> * built branch by_row_9ff1fc0f
#> * start branch by_row_c28ade2e
#> * built branch by_row_c28ade2e
#> * start branch by_row_bfe79c7e
#> * built branch by_row_bfe79c7e
#> * built pattern by_row
#> * end pipeline: 0.2 seconds
#> Warning messages:
#> 1: package 'targets' was built under R version 4.1.3 
#> 2: package 'dplyr' was built under R version 4.1.3 
#> 3: package 'sf' was built under R version 4.1.3 
tar_read(nc)
#>    AREA PERIMETER CNTY_ CNTY_ID        NAME  FIPS FIPSNO CRESS_ID BIR74 SID74
#> 1 0.114     1.442  1825    1825        Ashe 37009  37009        5  1091     1
#> 2 0.061     1.231  1827    1827   Alleghany 37005  37005        3   487     0
#> 3 0.143     1.630  1828    1828       Surry 37171  37171       86  3188     5
#> 4 0.070     2.968  1831    1831   Currituck 37053  37053       27   508     1
#> 5 0.153     2.206  1832    1832 Northampton 37131  37131       66  1421     9
#> 6 0.097     1.670  1833    1833    Hertford 37091  37091       46  1452     7
#>   NWBIR74 BIR79 SID79 NWBIR79
#> 1      10  1364     0      19
#> 2      10   542     3      12
#> 3     208  3616     6     260
#> 4     123   830     2     145
#> 5    1066  1606     3    1197
#> 6     954  1838     5    1237
#>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       geometry
#> 1                                                                                                                                                                                                                                        -81.47276, -81.54084, -81.56198, -81.63306, -81.74107, -81.69828, -81.70280, -81.67000, -81.34530, -81.34754, -81.32478, -81.31332, -81.26624, -81.26284, -81.24069, -81.23989, -81.26424, -81.32899, -81.36137, -81.36569, -81.35413, -81.36745, -81.40639, -81.41233, -81.43104, -81.45289, -81.47276, 36.23436, 36.27251, 36.27359, 36.34069, 36.39178, 36.47178, 36.51934, 36.58965, 36.57286, 36.53791, 36.51368, 36.48070, 36.43721, 36.40504, 36.37942, 36.36536, 36.35241, 36.36350, 36.35316, 36.33905, 36.29972, 36.27870, 36.28505, 36.26729, 36.26072, 36.23959, 36.23436
#> 2                                                                                                                                                                                                                                                             -81.23989, -81.24069, -81.26284, -81.26624, -81.31332, -81.32478, -81.34754, -81.34530, -80.90344, -80.93355, -80.96577, -80.94967, -80.95639, -80.97795, -80.98284, -81.00278, -81.02464, -81.04280, -81.08425, -81.09856, -81.11331, -81.12938, -81.13840, -81.15337, -81.17667, -81.23989, 36.36536, 36.37942, 36.40504, 36.43721, 36.48070, 36.51368, 36.53791, 36.57286, 36.56521, 36.49831, 36.46722, 36.41473, 36.40380, 36.39138, 36.37183, 36.36668, 36.37783, 36.41034, 36.42992, 36.43115, 36.42285, 36.42633, 36.41763, 36.42474, 36.41544, 36.36536
#> 3                                                                                                                                                                                                                   -80.45634, -80.47639, -80.53688, -80.54501, -80.55415, -80.59060, -80.62431, -80.66744, -80.69666, -80.72404, -80.73437, -80.75256, -80.76630, -80.78270, -80.87438, -80.87086, -80.88892, -80.92457, -80.95639, -80.94967, -80.96577, -80.93355, -80.90344, -80.83816, -80.61105, -80.43531, -80.45301, -80.45634, 36.24256, 36.25473, 36.25674, 36.27666, 36.27843, 36.26828, 36.27310, 36.24610, 36.25909, 36.25847, 36.26476, 36.25830, 36.26184, 36.24858, 36.23388, 36.32462, 36.35443, 36.37275, 36.40380, 36.41473, 36.46722, 36.49831, 36.56521, 36.56344, 36.55730, 36.55104, 36.25709, 36.24256
#> 4 -76.00897, -76.01735, -76.03288, -76.04395, -76.09509, -76.16093, -76.15815, -76.16829, -76.33025, -76.12740, -76.04596, -76.03321, -76.09106, -75.97607, -75.96976, -76.00161, -75.95126, -75.92812, -75.92459, -75.80006, -75.79885, -75.85516, -75.91376, -75.95751, -75.94193, -76.00897, 36.31960, 36.33773, 36.33598, 36.35359, 36.34892, 36.39190, 36.41269, 36.42709, 36.55606, 36.55716, 36.55695, 36.51437, 36.50357, 36.43621, 36.41512, 36.41891, 36.36547, 36.42324, 36.35095, 36.11282, 36.07282, 36.10567, 36.24480, 36.25945, 36.29434, 36.31960, -76.02717, -75.99866, -75.91192, -75.92480, -75.97728, -75.97629, -76.02717, 36.55672, 36.55665, 36.54253, 36.47398, 36.47802, 36.51793, 36.55672, -75.90199, -75.87817, -75.77316, -75.78317, -75.90199, 36.55620, 36.55587, 36.22926, 36.22519, 36.55620
#> 5                                                                                     -77.21767, -77.23461, -77.29861, -77.29351, -77.30948, -77.33499, -77.38903, -77.37862, -77.41347, -77.41885, -77.45412, -77.53808, -77.55743, -77.57195, -77.58008, -77.55968, -77.56042, -77.63597, -77.65099, -77.69888, -77.74940, -77.90121, -77.89886, -77.76393, -77.32005, -77.17735, -77.15442, -77.09213, -77.07531, -77.08327, -77.12733, -77.13932, -77.14196, -77.21767, 36.24098, 36.21460, 36.21153, 36.16286, 36.16277, 36.18285, 36.20310, 36.24009, 36.25582, 36.28180, 36.31976, 36.30246, 36.30421, 36.31450, 36.32827, 36.37595, 36.40636, 36.44054, 36.47259, 36.47896, 36.47357, 36.50989, 36.55294, 36.55344, 36.55392, 36.55629, 36.52625, 36.50752, 36.48352, 36.47010, 36.47071, 36.45648, 36.41706, 36.24098
#> 6                                                                                                                                                                                                                                                                                                                                                 -76.74506, -76.98069, -76.99475, -77.13007, -77.21767, -77.14196, -77.13932, -77.12733, -77.08327, -77.07531, -77.09213, -77.15442, -77.17735, -76.92413, -76.90842, -76.94577, -76.95367, -76.94351, -76.92408, -76.74135, -76.70750, -76.74506, 36.23392, 36.23024, 36.23558, 36.23346, 36.24098, 36.41706, 36.45648, 36.47071, 36.47010, 36.48352, 36.50752, 36.52625, 36.55629, 36.55415, 36.50428, 36.45896, 36.41923, 36.40173, 36.39244, 36.31517, 36.26613, 36.23392
tar_read(by_row)
#> Error in `stop_vctrs()`:
#> ! Input must be a vector, not a <sfc_MULTIPOLYGON/sfc> object.
library(sf)
#> Warning: package 'sf' was built under R version 4.1.3
#> Linking to GEOS 3.9.1, GDAL 3.2.1, PROJ 7.2.1; sf_use_s2() is TRUE
tar_read(by_row)
#> Simple feature collection with 6 features and 1 field
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -81.74107 ymin: 36.23436 xmax: -81.23989 ymax: 36.58965
#> Geodetic CRS:  NAD27
#>    AREA                       geometry
#> 1 0.114 MULTIPOLYGON (((-81.47276 3...
#> 2 0.061 MULTIPOLYGON (((-81.23989 3...
#> 3 0.143 MULTIPOLYGON (((-80.45634 3...
#> 4 0.070 MULTIPOLYGON (((-76.00897 3...
#> 5 0.153 MULTIPOLYGON (((-77.21767 3...
#> 6 0.097 MULTIPOLYGON (((-76.74506 3...

^{Created on 2022-03-18 by the reprex package (v2.0.1.9000)}

4 replies

wlandau Mar 18, 2022
Maintainer

Yeah, so apparently the sf package is designed so that you need library(sf) in order to work with sf objects. This situation is not very common, so targets only knows about the packages you need in order to run the target commands, as opposed to any extra packages needed to work with the data. For your project, I recommend creating a project-level .Rprofile file with library(sf) so the main process and any parallel workers will always have sf loaded.

hansvancalster Mar 21, 2022

Thanks, the .Rprofile workaround seems to work great!
Still, it remains a bit a mystery to me why calling tar_options_set(packages = "sf") does not work for the second tar_make() call in the reprex above. According to the help file of tar_options_set, packages = "Character vector of packages to load right before the target builds." I would therefore expect it to work also in the second call to tar_make() in the above reprex (which involves rebuilding a target due to a changed command). Is this a bug?

wlandau Mar 21, 2022
Maintainer

It's an edge case, not a bug. As I explained, the packages option was for running targets, not for loading upstream dependencies. And the packages are not loaded until the target runs. In that failing reprex, sf was needed to load nc, but tar_make() did not load any packages because no targets had been started up to that point. So targets was doing what I designed it to do even though that did not happen to work for this rare situation.

But in any case, that reprex should work after eeedddf because packages now applies to loading data for downstream targets as well as running the current target.

hansvancalster Mar 21, 2022

Thank you so much! 👍

Branch over sf objects #713

Uh oh!

Replies: 3 comments · 13 replies

Uh oh!

wlandau Nov 30, 2021 Maintainer

Uh oh!

wlandau Nov 30, 2021 Maintainer

Uh oh!

Uh oh!

lindsayplatt Nov 30, 2021 Author

Uh oh!

lindsayplatt Nov 30, 2021 Author

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

wlandau Mar 17, 2022 Maintainer

Uh oh!

Uh oh!

Uh oh!

wlandau Mar 18, 2022 Maintainer

Uh oh!

Uh oh!

wlandau Mar 21, 2022 Maintainer

Uh oh!

Replies: 3 comments 13 replies

wlandau
Nov 30, 2021
Maintainer

wlandau Nov 30, 2021
Maintainer

lindsayplatt Nov 30, 2021
Author

lindsayplatt Nov 30, 2021
Author

wlandau Mar 17, 2022
Maintainer

wlandau Mar 18, 2022
Maintainer

wlandau Mar 21, 2022
Maintainer