coops_search() time_zone weirdness?

[drf5n]

> ss= coops_search(begin_date=20161210, end_date=20161210,station_name=8638610,product='water_level',datum='NAVD',time_zone='lst')
> ss$data$t[2]
[1] "2016-12-10 00:06:00 EST"
> ss$data$v[2]
[1] -0.453
> ss= coops_search(begin_date=20161210, end_date=20161210,station_name=8638610,product='water_level',datum='NAVD',time_zone='gmt')
> ss$data$t[2]
[1] "2016-12-10 00:06:00 EST"
> ss$data$v[2]
[1] 0.019
> 

These are consistent with what you see at https://tidesandcurrents.noaa.gov/waterlevels.html?id=8638610&units=metric&bdate=20161210&edate=20161210&timezone=GMT&datum=NAVD&interval=6&action=data

I would expect the resultant values to be the same if the times are the same, but the vaues differ while the time remains the same.

If I unclass the resultant time, I see that the timezone attribute exists but is not set:

> unclass(ss$data$t[2])
[1] 1481346360
attr(,"tzone")
[1] ""

Here, I'd expect the attribute to be set appropriately, as GMT, and EST or EST/EDT for the Sewells Point Virginia station.

How should one use the resultant data to get consistent times?

[sckott]

thanks for the issue @drf5n

I'll have a look

[drf5n]

My workaround right now is to set the timezone to GMT, do the request, then assign GMT to the returned data:

Sys.setenv('TZ'='GMT')  # make sure the system time attribute is set.
Sys.timezone()
ss= coops_search(begin_date=20161210, end_date=20161210,station_name=8638610,product='water_level',datum='NAVD',time_zone='gmt')
attr(ss$data$t,'tzone')<-'GMT'  # fix up the time attribute

ss$data$t[2]
# [1] "2016-12-10 00:06:00 GMT"
Sys.setenv('TZ'='EST')  # change TZ to see if it affects the data... nope:
ss$data$t[2]
# [1] "2016-12-10 00:06:00 GMT"

[sckott]

one problem is that if we need to use LST, that's the local time zone of the station, but they don't give us the time zone of the station in the response. so we have to get it somehow,

then i can correct the date values given back

[sckott]

@drf5n do you ever use lst_ldt?

[sckott]

try again after reinstalling

devtools::install_github("ropensci/rnoaa")
library(rnoaa)
ss <- coops_search(begin_date=20161210, end_date=20161210,station_name=8638610,product='water_level',datum='NAVD',time_zone='lst')
ss$data$t[2]
#> [1] "2016-12-10 00:06:00 EST"
rr <- coops_search(begin_date=20161210, end_date=20161210,station_name=8638610,product='water_level',datum='NAVD',time_zone='gmt')
rr$data$t[2]
#> [1] "2016-12-10 00:06:00 GMT"

head(ss$data, n = 3)
#>                     t      v     s       f q
#> 1 2016-12-10 00:00:00 -0.458 0.005 0,0,0,0 v
#> 2 2016-12-10 00:06:00 -0.453 0.005 0,0,0,0 v
#> 3 2016-12-10 00:12:00 -0.444 0.004 0,0,0,0 v
rr$data[51:53,]
#>                      t      v     s       f q
#> 51 2016-12-10 05:00:00 -0.458 0.005 0,0,0,0 v
#> 52 2016-12-10 05:06:00 -0.453 0.005 0,0,0,0 v
#> 53 2016-12-10 05:12:00 -0.444 0.004 0,0,0,0 v

# convert
ss$data$t <- as.POSIXlt(ss$data$t, tz = "GMT")

head(ss$data, n = 3)
#>                     t      v     s       f q
#> 1 2016-12-10 05:00:00 -0.458 0.005 0,0,0,0 v
#> 2 2016-12-10 05:06:00 -0.453 0.005 0,0,0,0 v
#> 3 2016-12-10 05:12:00 -0.444 0.004 0,0,0,0 v
rr$data[51:53,]
#>                      t      v     s       f q
#> 51 2016-12-10 05:00:00 -0.458 0.005 0,0,0,0 v
#> 52 2016-12-10 05:06:00 -0.453 0.005 0,0,0,0 v
#> 53 2016-12-10 05:12:00 -0.444 0.004 0,0,0,0 v

[sckott]

changes in ddad05b

[drf5n]

Re. LST_LDT, no, I try to always use UTC.

This fix works for me. Thanks.

[sckott]

Glad it works

[sckott]

done