I'm learning the sf package in R and would like to locate about 450 sites along the nearest reaches of a river network and to obtain the upstream and downstream distances along the reach where the site is located. I've include one simple feature POINT (site) and LINESTRING or polyline (reach) to provide a reproducible example at the bottom of my post.

Given a point located exactly along the polyline, how would I split the polyline at that point to obtain two polylines? I would like to obtain the lengths of the split polylines both upstream and downstream of the point in order to edit a graph representation of the whole river.

A similar question on SE is submitted here but involves splitting a polyline using polygons. My latest attempt with st_split() yields a GEOMETRYCOLLECTION, but it seems that the LINESTRING coordinates are not listed separately:

# Get the site/reach objects from dput()...
split.reach <- st_split(reach$geometry, site$geometry)
> split.reach
Geometry set for 1 feature 
dimension:      XY
bbox:           xmin: 688358.8 ymin: 171876.4 xmax: 688465 ymax: 172171.2
epsg (SRID):    NA
proj4string:    +proj=somerc +lat_0=46.95240555555556 +lon_0=7.439583333333333 +k_0=1 +x_0=600000 +y_0=200000 +ellps=bessel +units=m +no_defs

To be clear, in this specific problem I'm interested in lengths along an individual river reach (i.e., an individual row or LINESTRING in an 'sf' object) that is split by a point and not along the whole river network. I would also like to avoid writing code that depends on a large number of spatial packages (i.e., stick with sf)

The dput for a site and reach as an example:

site <- structure(list(SiteId = "PF_00183", SampId = "PF_00183 _ 2014-10-15", 
    Longitude.In = 8.59713558862371, Latitude.In = 46.6953718769341, 
    X.In = 688524.841170469, Y.In = 172082.839248429, ReachId = 125010L, 
    Distance = 76.8220253856169, Longitude.Near = 8.59613333686445, 
    Latitude.Near = 46.6954168201561, geometry = structure(list(
        structure(c(688448.125, 172086.703125001), class = c("XY", 
        "POINT", "sfg"))), n_empty = 0L, precision = 0, crs = structure(list(
        epsg = NA_integer_, proj4string = "+proj=somerc +lat_0=46.95240555555556 +lon_0=7.439583333333333 +k_0=1 +x_0=600000 +y_0=200000 +ellps=bessel +units=m +no_defs"), .Names = c("epsg", 
    "proj4string"), class = "crs"), bbox = structure(c(688448.125, 
    172086.703125001, 688448.125, 172086.703125001), .Names = c("xmin", 
    "ymin", "xmax", "ymax"), class = "bbox"), class = c("sfc_POINT", 
    "sfc"))), .Names = c("SiteId", "SampId", "Longitude.In", 
"Latitude.In", "X.In", "Y.In", "ReachId", "Distance", "Longitude.Near", 
"Latitude.Near", "geometry"), row.names = c(NA, -1L), sf_column = "geometry", agr = structure(c(NA_integer_, 
NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
NA_integer_, NA_integer_, NA_integer_, NA_integer_), class = "factor", .Label = c("constant", 
"aggregate", "identity"), .Names = c("SiteId", "SampId", "Longitude.In", 
"Latitude.In", "X.In", "Y.In", "ReachId", "Distance", "Longitude.Near", 
"Latitude.Near")), class = c("sf", "data.table", "data.frame"
reach <- structure(list(FNODE_ = 148558, TNODE_ = 148142, LENGTH = 343.00790968, 
    ReachId = 125010L, geometry = structure(list(structure(c(688400.125, 
    688394.375, 688385.875, 688375.6875, 688363.625, 688359.8125, 
    688358.8125, 688361.125, 688375.6875, 688389.375, 688411.1875, 
    688432.875, 688445.1875, 688448.125, 688445.875, 688446.625, 
    688448.125, 688455.375, 688465, 171876.40625, 171890.40625, 
    171906.203125, 171916.703125, 171929.09375, 171944.703125, 
    171952.59375, 171959.90625, 171980.296875, 171997.09375, 
    172017.796875, 172041.90625, 172063.5, 172086.703125, 172109.90625, 
    172129, 172138.5, 172153.5, 172171.203125), .Dim = c(19L, 
    2L), class = c("XY", "LINESTRING", "sfg"))), class = c("sfc_LINESTRING", 
    "sfc"), precision = 0, bbox = structure(c(688358.8125, 171876.40625, 
    688465, 172171.203125), .Names = c("xmin", "ymin", "xmax", 
    "ymax"), class = "bbox"), crs = structure(list(epsg = NA_integer_, 
        proj4string = "+proj=somerc +lat_0=46.95240555555556 +lon_0=7.439583333333333 +k_0=1 +x_0=600000 +y_0=200000 +ellps=bessel +units=m +no_defs"), .Names = c("epsg", 
    "proj4string"), class = "crs"), n_empty = 0L)), .Names = c("FNODE_", 
"TNODE_", "LENGTH", "ReachId", "geometry"), row.names = 125010L, class = c("sf", 
"data.frame"), sf_column = "geometry", agr = structure(c(NA_integer_, 
NA_integer_, NA_integer_, NA_integer_), .Names = c("FNODE_", 
"TNODE_", "LENGTH", "ReachId"), .Label = c("constant", "aggregate", 
"identity"), class = "factor"))
  • the site object you provide has no crs set. Can you update? Jul 4, 2018 at 14:54
  • Thank you for pointing that out. I have updated the object.
    – bcaradima
    Jul 4, 2018 at 15:03

1 Answer 1


If you snap the point to the line then you can split the line and then extract the parts from the resulting collection. Use a really small tolerance, I don't know how small it needs to be...

> site_snap = st_snap(site, reach, tol=1e-9)
> parts = st_collection_extract(st_split(reach$geometry, site_snap$geometry),"LINESTRING")

parts is now two features, either side of your point:

> parts
Geometry set for 2 features 
geometry type:  LINESTRING
dimension:      XY
bbox:           xmin: 688358.8 ymin: 171876.4 xmax: 688465 ymax: 172171.2
epsg (SRID):    NA
proj4string:    +proj=somerc +lat_0=46.95240555555556 +lon_0=7.439583333333333 +k_0=1 +x_0=600000 +y_0=200000 +ellps=bessel +units=m +no_defs
LINESTRING (688400.1 171876.4, 688394.4 171890....
LINESTRING (688448.1 172086.7, 688445.9 172109....

Your original site doesn't intersect the line, so your split returned nothing. Test with:

> st_intersects(site, reach, sparse=FALSE)
[1,] FALSE
> st_intersects(site_snap, reach, sparse=FALSE)
[1,] TRUE

Your original site is 10^-11 units from the line feature:

> st_distance(site$geom, reach$geom)
Units: m
[1,] 9.550647e-11
  • Thanks, this is clearly a working solution. But, I don't really understand how the tolerance argument for st_snap actually works, and the documentation for sf and PostGIS don't really explain it either. Could st_snap also work if the point is, say, 75m away from the reach?
    – bcaradima
    Jul 5, 2018 at 11:37
  • I think as it stands this post has two questions, and I've only answered the second one (how to cut a line by points on the line). The question of how to get a point on a line is a separate question, and you should probably edit that much out of this post and start a new post which is specific to that aspect.
    – Spacedman
    Jul 5, 2018 at 13:32
  • Thanks Spacedman, I have posted a new question here: gis.stackexchange.com/questions/288570/…
    – bcaradima
    Jul 5, 2018 at 15:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.