Polars

Misc

• Also see Python, Polars

• Packages

  • {polars}
  • {tidypolars} - Allows one to use the tidyverse syntax while using the power of polars
  • {polarssql} - Provides a polars backend for DBI and dbplyr.

• Resources

  • Cookbook Polars for R
  • Polars from Python and R

• as.vector(<series>) will remove attributes that might be useful. Therefore, recommended to use <series>$to_r_vector() instead for usual conversions

• Coerce an existing dataframe into a polars dataframe: as_polars_df or as_polars_lf for a lazy dataframe. For creating a df inside the DataFrame, you need !!!

  pl$DataFrame(!!!data.frame(x = 1, y = "a"))
  #> shape: (1, 2)
  #> ┌─────┬─────┐
  #> │ x   ┆ y   │
  #> │ --- ┆ --- │
  #> │ f64 ┆ str │
  #> ╞═════╪═════╡
  #> │ 1.0 ┆ a   │
  #> └─────┴─────┘

• if you used to pass a vector of column names or a list of expressions, you need to expand it with !!!

  dat <- as_polars_df(head(mtcars, 3))
  my_exprs <- list(pl$col("drat") + 1, "mpg", "cyl")
  dat$select(!!!my_exprs)
  #> shape: (3, 3)
  #> ┌──────┬──────┬─────┐
  #> │ drat ┆ mpg  ┆ cyl │
  #> │ ---  ┆ ---  ┆ --- │
  #> │ f64  ┆ f64  ┆ f64 │
  #> ╞══════╪══════╪═════╡
  #> │ 4.9  ┆ 21.0 ┆ 6.0 │
  #> │ 4.9  ┆ 21.0 ┆ 6.0 │
  #> │ 4.85 ┆ 22.8 ┆ 4.0 │
  #> └──────┴──────┴─────┘
  
  pl$col(!!!c("foo", "bar"), "baz")
  #> cols(["foo", "bar", "baz"])

  • Same with lazyframe>$cast() and <lazyframe>$group_by()

Mutate

• Example: 10-day and 50-day Moving Average (source)

  moving_average_pl <- 
    long_pl$with_columns(
      pl$
        col("Price")$
        rolling_mean(10)$
        over("Stock")$
        alias("Price_MA10"),
      pl$
        col("Price")$
        rolling_mean(50)$
        over("Stock")$
        alias("Price_MA50")
  )
  moving_average_pl
  
  moving_average_pl |> 
      as_tibble() |> 
      tidyr::pivot_longer(
          cols = Price:Price_MA50
      ) %>%
      group_by(Stock) |> 
      plot_time_series(as_date(Date), value, .color_var = name, .facet_ncol = 4, .smooth = FALSE)

Summarize

• Example: min, max, mean by group

  df <- pl$scan_csv(file_name)$
      group_by("state")$
      agg(
          pl$
            col("measurement")$
            min()$
            alias("min_m"),
          pl$
            col("measurement")$
            max()$
            alias("max_m"),
          pl$
            col("measurement")$
            mean()$
            alias("mean_m")
      )$
      collect()

Pivoting

• Example: pivot_longer (source)

  long_pl = stock_data_pl$unpivot(
      index         = "Date",
      value_name    = "Price",
      variable_name = "Stock"
  )
  long_pl
  
  long_pl %>%
      as_tibble() |> 
      group_by(Stock)  |> 
      timetk::plot_time_series(as_date(Date), Price, .facet_ncol = 4, .smooth = FALSE)

SQL

• Example: min, max, mean by group

  lf <- pl$LazyFrame(D) 
  
  pl$
    SQLContext(frame = lf)$
    execute(
      "select min(measurement) as min_m, 
              max(measurement) as max_m, 
              avg(measurement) as mean_m 
      from frame 
      group by state"
    )$
    collect()