ClaytonJY

df <- data.frame(
  id    = 1:5,
  label = LETTERS[1:5],  # A, B, C, D, E
  gen_1 = c(1.0, 1.1, NA, 1.3, NA),
  gen_2 = c(2.0, NA, 2.2, 2.3, NA),
  gen_3 = c(3.1, NA, 3.2, NA, NA)
)

# take a look
df

ClaytonJY

library(tidyverse)

# fake some data
input_tbl <- lst(
  MarketDay = as.Date(c("2016-01-01", "2016-01-02", "2016-01-03")),
  HourEnding = 1:3,
  FuelType = c("Coal", "Gas", "Hydro"),
  Month = 1,  # because only Jan dates
  Region = "North"

ClaytonJY

# built-in function for "today"
Sys.Date()
#> [1] "2019-11-11"

# Dates are actually just integers underneath
# we can coerce to integer to see that version of it
today_days <- as.integer(Sys.Date())
today_days
#&gt; [1] 18211

ClaytonJY

Timescale Selection Query Plans

I've seen some odd behavior trying to select specific items from Timescale hypertables, which I reproduce here with randomized data.

Setup

We'll need a vanilla Timescale instance; I like Docker.

# start a PG12 instance w/ Timescale 2.0