From a7f0c90b4638690ce589522bfcc99347172552fd Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bogumi=C5=82=20Kami=C5=84ski?= <bkamins@sgh.waw.pl>
Date: Mon, 14 Feb 2022 16:37:49 +0100
Subject: [PATCH] add chapter 11

---
 appB.jl |  6 +++++
 ch11.jl | 75 ++++++++++++++++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 78 insertions(+), 3 deletions(-)

diff --git a/appB.jl b/appB.jl
index 42c4ecd..5b0755d 100644
--- a/appB.jl
+++ b/appB.jl
@@ -165,6 +165,12 @@ proptable([walk_unique_2ahead() for _ in 1:10^5])
 @time wide = DataFrame(ones(1, 10_000), :auto);
 @time Tables.columntable(wide);
 
+# Code for exercise 11.2
+
+using Statistics
+Dict(key.city => mean(df.rainfall) for (key, df) in pairs(gdf_city))
+combine(gdf_city, :rainfall => mean)
+
 # Code for exercise 12.1
 
 cg = complete_graph(37700)
diff --git a/ch11.jl b/ch11.jl
index 565b295..24cc580 100644
--- a/ch11.jl
+++ b/ch11.jl
@@ -4,6 +4,12 @@
 
 # Code for section 11.1
 
+# deserialization of source data frame
+
+using DataFrames
+using Serialization
+walk = deserialize("walk.bin")
+
 # Code for a note on conversion
 
 x = [1.5]
@@ -12,9 +18,6 @@ x
 
 # Code from section 11.1.1
 
-using Serialization
-walk = deserialize("walk.bin")
-
 Matrix(walk)
 Matrix{Any}(walk)
 Matrix{String}(walk)
@@ -83,3 +86,69 @@ identity.(eachcol(walk))
 
 df = DataFrame(x=1:2, b=["a", "b"])
 identity.(eachcol(df))
+
+# Code from section 11.2
+
+using CSV
+raw_data = """
+city,date,rainfall
+Olecko,2020-11-16,2.9
+Olecko,2020-11-17,4.1
+Olecko,2020-11-19,4.3
+Olecko,2020-11-20,2.0
+Olecko,2020-11-21,0.6
+Olecko,2020-11-22,1.0
+Ełk,2020-11-16,3.9
+Ełk,2020-11-19,1.2
+Ełk,2020-11-20,2.0
+Ełk,2020-11-22,2.0
+""";
+rainfall_df = CSV.read(IOBuffer(raw_data), DataFrame)
+
+gdf_city = groupby(rainfall_df, "city")
+
+gdf_city_date = groupby(rainfall_df, Not("rainfall"))
+
+keys(gdf_city_date)
+
+gk1 = keys(gdf_city_date)[1]
+g1_t = Tuple(gk1)
+g1_nt = NamedTuple(gk1)
+g1_dict = Dict(gk1)
+
+gdf_city_date[1]
+gdf_city_date[gk1]
+gdf_city_date[g1_t]
+gdf_city_date[g1_nt]
+gdf_city_date[g1_dict]
+
+gdf_city[("Olecko",)]
+gdf_city[(city="Olecko",)]
+
+using BenchmarkTools
+bench_df = DataFrame(id=1:10^8);
+bench_gdf = groupby(bench_df, :id);
+@btime groupby($bench_df, :id);
+bench_i = 1_000_000;
+bench_gk = keys(bench_gdf)[bench_i];
+bench_t = Tuple(bench_gk);
+bench_nt = NamedTuple(bench_gk);
+bench_dict = Dict(bench_gk);
+@btime $bench_gdf[$bench_i];
+@btime $bench_gdf[$bench_gk];
+@btime $bench_gdf[$bench_t];
+@btime $bench_gdf[$bench_nt];
+@btime $bench_gdf[$bench_dict];
+
+gdf_city[[2, 1]]
+gdf_city[[1]]
+
+[nrow(df) for df in gdf_city]
+
+for p in pairs(gdf_city)
+    println(p)
+end
+
+Dict(key.city => nrow(df) for (key, df) in pairs(gdf_city))
+
+combine(gdf_city, nrow)