hadoop - Pig - 针对主表映射和检索两列？

Question

我正在 openflights 数据集 ( https://openflights.org/data.html ) 上试验 pig。我目前正在尝试映射一个包含所有唯一可能航线的查询，即下表

+---------------+-------------+
| Start_Airport | End_Airport |
+---------------+-------------+
| YYZ           | NYC         |
| YBG           | YVR         |
| AEY           | GOH         |
+---------------+-------------+ 

然后将这两个值与包含每个机场的经度和纬度的主表连接起来。 IE。

+---------+----------+-----------+
| Airport | Latitude | Longitude |
+---------+----------+-----------+
| YYZ     |    -10.3 |      1.23 |
| YBG     |    -40.3 |      50.4 |
| AEY     |     30.3 |      30.3 |
+---------+----------+-----------+

我将如何尝试这样做？我本质上是想拥有一个看起来像的决赛 table

+----------------+----------+-----------+-------------+----------+-----------+
| Start_Airport  | Latitude | Longitude | End_Airport | Latitude | Longitude |
+----------------+----------+-----------+-------------+----------+-----------+
| YYZ            |    -10.3 |      1.23 | NYC         | blah     | blah      |
| YBG            |    -40.3 |      50.4 | YVR         | blah     | blah      |
| AEY            |     30.3 |      30.3 | GOH         | blah     | blah      |
+----------------+----------+-----------+-------------+----------+-----------+

我目前正在尝试执行以下操作，其中 c 是第一个表

route_data = JOIN c by (start_airport, end_airport), airports_all by ($0, $0);

我认为这基本上是针对查询说的，将starting_aiport和ending_airport与各自的代码结合起来，然后通过各自的经度和纬度，

Answer 1

route_data = JOIN c by (start_airport, end_airport), airports_all by ($0, $0);

这类似于 sql 世界中典型连接查询的“and”条件子句。想象一下下面的查询。它会产生你想要的结果。
select * from c t1 join airports_all t2 on a.start_airport=b.first_field 和 a.end_airport=b.first_field;只有当 start_airport 和 end_airport 相同时，才会产生结果。

您想要的可以通过以下方式实现:

cat > routes.txt
YYZ,NYC
YBG,YVR
AEY,GOH

cat > airports_all.txt
YYZ,-10.3,1.23
YBG,-40.3,50.4
AEY,30.3,30.3

pig 代码:

tab1 = load '/home/ec2-user/routes.txt' using PigStorage(',') as (start_airport,end_airport);
describe tab1
tab2 = load '/home/ec2-user/airports_all.txt' using PigStorage(',') as (Airport,Latitude,Longitude);
describe tab2
tab3 = JOIN tab1 by (start_airport), tab2 by (Airport);
describe tab3
tab4 = foreach tab3 generate $0 as start_airport, $3 as start_Latitude, $4 as start_Longitude, $1 as end_airport;
describe tab4
tab5 = JOIN tab4 by (end_airport), tab2 by (Airport);
describe tab5
tab6 = foreach tab5 generate $0 as start_airport, $1 as start_Latitude, $2 as start_Longitude, $3 as end_airport, $5 as end_Latitude, $6 as end_Longitude;
describe tab6
dump tab6