dplr_ideas.md

    df_data <- 
      tbl(src_azure, "devicestate") %>%
      select(dtm = UTC_DateTime, ActiveCool, Temperature_C, Sum_Power_wh) %>%
      filter(DeviceID == input$device) %>%
      filter(UTC_DateTime > dtm_range[[1]] & UTC_DateTime < dtm_range[[2]]) %>%
      collect()

The line filter(UTC_DateTime > dtm_range[[1]] & UTC_DateTime < dtm_range[[2]]) should work as filter(dtm > dtm_range[[1]] & dtm < dtm_range[[2]])

native support for date-times?

MySQL does not handle %in% for singular values

towards standard evaluation

let's say we want to specify which columns to select

a <- "col_name"
df %>% select(get(a))

attributes

data.frames with columns that have attributes (outside of POSIXct) choke the group_by function

ggvis_idea.md

list of things to propose/discuss for ggvis vignettes

• hide_axis
• hide_legend
• expand
• the size options refer to the grid size, not the plot size (?)
• make reactive only what you need (don't wrap the whole darn thing)
• support for images

need to make reproducible example for

• disappearing histogram when duration != 0 (solution might have something to do with setting key)
• using reactive data, why does image corect only after resizing (wrong color, for example) - again, this goes away when duration == 0

use case for producing an empty ggvis when data is empty