Extract standard errors from Arima model applied to groups using sweep I am following the sweep vignette on Forecasting Time Series Groups in the tidyverse, see here. sweep is the broom package for time series data.
I am estimating an AR(1) model using Arima from the forecasting package on each group. I can estimate the AR(1) for each group, and I can report the coefficients tidily with sw_tidy. However I cannot report the standard errors that correspond to the coefficients tidily. 
ind_ts is a dataframe where rows are groups and there is a column data_ts. An element of the column data_ts is a list of time series data. I use the following command to apply the Arima model to the time-series data in each row of data_ts:
ind_ar <- ind_ts %>%
  mutate(fit.ar = map(data_ts, Arima))

The column fit.ar contains the output of the AR(1) model applied to each group's time series. To get the coefficients of this AR(1) in tidy format I use sw_tidy:
ind_ar %>%
  mutate(tidy = map(fit.ar, sw_tidy)) %>%
  unnest(tidy) 

However I would also like a column with the corresponding standard errors for estimates i.e. ind_ar$std.er.
 A: Tidy forecasting is now much easier using the fable package rather than a combination of forecast and sweep. Here is an example.
library(fable)
#> Loading required package: fabletools
library(tidyverse)


ind_ts <- tibble(X = rnorm(20), Y=rnorm(20), time=1:20) %>%
  gather(-time, key="Series", value="value") %>%
  as_tsibble(index=time, key=Series)

ind_ar <- ind_ts %>%
  model(fit_ar = ARIMA(value ~ pdq(p=1,d=0,q=0)))


ind_ar %>% tidy()
#> # A tibble: 2 x 7
#>   Series .model term  estimate std.error statistic p.value
#>   <chr>  <chr>  <chr>    <dbl>     <dbl>     <dbl>   <dbl>
#> 1 X      fit_ar ar1    -0.0199     0.219   -0.0909   0.928
#> 2 Y      fit_ar ar1     0.176      0.219    0.806    0.430

Created on 2019-10-31 by the reprex package (v0.3.0)
A: At the moment I obtain them by digging into the var.coef method. This reports the variance-covariance matrix and I take the square root of the main diagonal to return standard errors.
vcov <- lapply(ind_ar$fit.ar, `[[`, 'var.coef')

std.er = lapply(
    vcov, 
    function(x) 
    {
      y <- sqrt(diag(x))
      y
    }
  )

However this does not add them as a column corresponding to their estimates.
Given it is typical to report coefficient estimates and standard errors, I expect there is a more elegant way than mine to do this.
