Cheatsheet

scala-cli repl --dep io.github.quafadas::scautable::0.0.34 \
  --scalac-option -Xmax-inlines --scalac-option 2048 \
  --java-opt -Xss4m \
  --repl-init-script 'import io.github.quafadas.table.{*, given}'

Reading CSV Files

Csv Available As	CSV Size	Columns	Hints
Inlined in code	<1Kb	< 100	`inline val s= "aHeader\nr1c1\nr2c1"` `CSV.fromString(s)`
File - Compile Time	< 20Mb	<100	`CSV.pwd("a.csv")` `CSV.resource("a.csv")` `CSV.absolutePath("/abs/path/to/a.csv")`
File - Compile Time	> 20Mb & < 250 Mb		Dont attempt to infer types at compile time. `val inferOpts = CsvOpts(TypeInferrer.FirstN(20000))` `val knownOpts = CsvOpts(TypeInferrer.FromTuple[(Int, String, Double)])` `CSV.pwd("/path/a.csv", inferOpts)` `CSV.absolutePath("/path/a.csv", knownOpts)` `CSV.resource("a.csv", knownOpts)`
File - Run Time	< 250Mb	<100	`val reader = CSV.fromTyped[("col1", "col2", "col3"), (String, Int, Double)]` `val data = reader(os.pwd / "simple.csv")`
File	> 250 Mb or	>100 columns	Scautable is not the right thing to analyse this file. Consider a proper query optimised dataframe library such as Spark or Polars

Displaying Tables

Assuming you have an Iterator / Iterable of named tuples.

e.g. val data : Seq[(col1 : String, col2 : Int, col3 : Double)] = ???

Target	Hints
repl print / markdown	`data.ptbln`
String	`data.consoleFormatNt(fansi = false)`
html	`HtmlRenderer.nt(data)`
Almond	`Html(HtmlRenderer.nt(data))`
browser window	`HtmlRenderer.desktopShowNt(data)`

Excel Operations (JVM only)

TBD

Type Inference

Inferrer	Use Case
`FromAllRows`	Most accurate type detection (default)
`FirstRow`	Fast, but less accurate type detection
`StringType`	Safe Mode
`FromFirstNRows(n)`	Balance between accuracy and speed
`FromTuple[T]`	You have control of type inference. Can decode complex types via implicit evidence