ex3/main_3.tex

   1 %ex3.3
   2
   3 \begin{enumerate}[label=(\alph*),leftmargin=2\parindent]
   4         \item\textbf{Two basic queries to start with:}
   5         \begin{itemize}
   6                 \item\textbf{List all the distinct countries for which addresses are
   7                         registered in the database.}
   8                         \begin{lstlisting}[style=command]
   9 MATCH (n:Address) RETURN DISTINCT n.countries AS countries
  10                         \end{lstlisting}
  11                         This returns 222 different countries. We chose ''Austria''.
  12                         \newline
  13                         \newline
  14                 \item\textbf{Choose one of these countries and list 5 entities which
  15                         are based in that country.}
  16                         \begin{lstlisting}[style=command]
  17 MATCH (n:Address {countries: 'Austria'})<-[r:REGISTERED_ADDRESS]-(e:Entity)
  18 RETURN DISTINCT e AS entity LIMIT 5
  19                         \end{lstlisting}
  20                         \input{results/result_3a.tex}
  21                         \newline
  22                         \newline
  23         \end{itemize}
  24         \item
  25                 \begin{enumerate}[label=\roman*,leftmargin=1\parindent]
  26                         \item\textbf{First, find the 10 \texttt{top intermediaries} , i.e.
  27                                 those 10 intermediaries that have the most outgoing
  28                                 \texttt{INTERMEDIARY\_OF} edges. Output the name of the
  29                                 intermediary as well as the number of relevant edges.}
  30                                 \begin{lstlisting}[style=command]
  31 MATCH (n:Intermediary)-[r:INTERMEDIARY_OF]->()
  32 RETURN n.name AS name, COUNT(r) AS count ORDER BY count DESC LIMIT 10
  33                                 \end{lstlisting}
  34                                 \input{results/result_3b1.tex}
  35                                 \newline
  36                                 \newline
  37                         \item\textbf{Extend the query of \texttt{i} to also count outgoing
  38                                 \texttt{OFFICER\_OF} edges to determine the top intermediaries.
  39                                 Output the name of the intermediary as well as the number of
  40                                 relevant edges.}
  41                                 \begin{lstlisting}[style=command]
  42 MATCH (n:Intermediary)-[r:INTERMEDIARY_OF|OFFICER_OF]->()
  43 RETURN n.name AS name, COUNT(r) AS count ORDER BY count DESC LIMIT 10
  44                                 \end{lstlisting}
  45                                 \input{results/result_3b2.tex}
  46                                 \newline
  47                                 \newline
  48                         \item\textbf{For the top intermediary from query \texttt{ii},
  49                                 output all the outgoing edges, except for those that have type
  50                                 either \texttt{OFFICER\_OF} or \texttt{INTERMEDIARY\_OF}. Also
  51                                 output the respective nodes at the other side of the edges.}
  52                                 \begin{lstlisting}[style=command]
  53 MATCH (n:Intermediary {node_id: "54662"})-[r]->(t)
  54         WHERE NOT (n)-[r:INTERMEDIARY_OF|OFFICER_OF]->(t)
  55 RETURN type(r) AS relation, t.node_id AS node
  56                                 \end{lstlisting}
  57                                 \input{results/result_3b3.tex}
  58                                 \newline
  59                                 \newline
  60                 \end{enumerate}
  61         \item\textbf{Find a shortest \textit{undirected} path between an Address in
  62                 Luxembourg and an Address in Cyprus where the path has a length of at
  63                 least 16 and at most 30.}
  64
  65                 \begin{lstlisting}[style=command]
  66 MATCH p=shortestPath((lux:Address{countries:"Luxembourg"})-[*1..30]-(cyp:Address{countries:"Cyprus"}))
  67 WITH *, length(p) AS len WHERE len >= 16 AND len < 17
  68 //RETURN lux, cyp, len LIMIT 10
  69 RETURN p LIMIT 1
  70                 \end{lstlisting}
  71
  72                 Adjusting the values only produced different graphs, but there were
  73                 always more than 10 paths available.
  74                 One of these graphs we included, to show at least one result.
  75
  76                 \input{results/result_3c.tex}
  77
  78                 \begin{figure}[H]
  79                         \centering
  80                         \includegraphics[width=12cm]{results/result_3c.png}
  81                 \end{figure}
  82         \item\textbf{Find one (use \texttt{LIMIT 1}) subgraph of the form specied
  83                 in Figure 4 in the database.
  84                 Furthermore, the two entities should be in different countries and at
  85                 least one of the entities should have different values in its
  86                 \texttt{jurisdiction} and \texttt{country\_codes} attributes.}
  87                 \begin{lstlisting}[style=command]
  88 MATCH (i:Intermediary)-->(e1:Entity)<--(o:Officer)-->(e2:Entity)<--(i:Intermediary)
  89         WHERE e1 <> e2
  90         AND e1.countries <> e2.countries
  91         AND (e1.jurisdiction <> e1.country_codes
  92                 OR e2.jurisdiction <> e2.country_codes)
  93 RETURN i, e1, e2, o
  94 LIMIT 1
  95                 \end{lstlisting}
  96                 \begin{figure}[H]
  97                         \centering
  98                         \includegraphics[width=6cm]{results/result_3d.png}
  99                 \end{figure}
 100         \item\textbf{For the final task we want to make the \texttt{countries}
 101                 attribute a real part of the graph.
 102                 The best way to do this is in Neo4j is using the \texttt{MERGE} keyword
 103                 inside of a \texttt{MATCH}. The task is restricted to the nodes with
 104                 label \texttt{Other} to avoid processing an unnecessarily large amount
 105                 of data.}
 106                 \begin{lstlisting}[style=command]
 107 MATCH (other:Other) WHERE exists(other.countries)
 108 MERGE (country:Country { name: other.countries })
 109 MERGE (other)-[r:IN_COUNTRY]->(country)
 110                 \end{lstlisting}
 111                 \input{results/result_3e.tex}
 112                 \begin{lstlisting}[style=command]
 113 MATCH (c:Country{ name:"Mauritius" }) RETURN c
 114                 \end{lstlisting}
 115                 \begin{figure}[H]
 116                         \centering
 117                         \includegraphics[width=5cm]{results/result_3e.png}
 118                 \end{figure}
 119 \end{enumerate}