Skip to content

Latest commit

 

History

History
697 lines (628 loc) · 26.9 KB

2DArrayInQ.org

File metadata and controls

697 lines (628 loc) · 26.9 KB
Raw

D4M in Q

Library code

/Matrix multiplication of (D2D) Dictionary 2D.
.q.mm:{cc: cols[value x] inter cols y; flip[ y cc] {sum x*y}/:\: x[;cc]}
//For better display, change 0 to null.
Nullify:{@[x; where x=0; :; x`]} //here we assume the key tybe is symbol, and null symbol map to null value.
/ remove null from a dictionary
RmNull: {where[not null x]#x}

Foundation: 1D Q Dictionary Arithmetic

When do math(+,-,*,/, any binary function) with dictionary, missing keys are filled, and values on common keys will do the corresponding math.

-1"a";
show a: `a`b!1 2
-1"\nb";
show b:   `b`c!3 4
-1"\n a+b:";
show a+b
-1"\na*b";
show a*b
-1"\na&b";
show a & b
-1"\na|b";
show a|b
a
a| 1
b| 2

b
b| 3
c| 4

 a+b:
a| 1
b| 5
c| 4

a*b
a| 1
b| 6
c| 4

a&b
a| 1
b| 2
c| 4

a|b
a| 1
b| 3
c| 4

Dictionary 2D(D2D) in Q

A dictionary from array to table. It can be flipped, and you can select rows, columns from it.

show t: `a`b!`u`x`y!/:v: 2 3# 1 1 0 0 0 1.0
show flip t
 | u x y
-| -----
a| 1 1 0
b| 0 0 1
 | a b
-| ---
u| 1 0
x| 1 0
y| 0 1
1 1 0
0 0 1

selection and assignment

-1@"t";
show t
-1"\nselect row: t[`a]";
show t[`a]
-1"\nselect columns: t[;`x`y] or .[t; :: ; `x`y]";

show t[;`x`y]
show .[t; (::;`x`y)]
-1"\nSelect multiple rows and columns: .[t; (`a`b;`x`y)]";
show t[`a`b;`x`y]
-1"\nselect cell: t[`a;`u]";
show t[`a;`u]
-1"\n.[t; `a`x; :; 3.0] //assign value to a cell";
show .[t; `a`x; :; 3.0] 

-1 "\nAssign row: .[t; (`a; ::); :; 3 2 1.0] ";
show .[t; (`a;::) ; :; 3 2 1.0] 
-1"\n Assign to columns: .[t;(::;`x`y); :; 9.0]";
show .[t; (::;`x`y); :; 9.0 ]
t
 | u x y
-| -----
a| 1 1 0
b| 0 0 1

select row: t[`a]
u| 1
x| 1
y| 0

select columns: t[;`x`y] or .[t; :: ; `x`y]
a| 1 0
b| 0 1
a| 1 0
b| 0 1

Select multiple rows and columns: .[t; (`a`b;`x`y)]
1 0
0 1

select cell: t[`a;`u]
1f

.[t; `a`x; :; 3.0] //assign value to a cell
 | u x y
-| -----
a| 1 3 0
b| 0 0 1

Assign row: .[t; (`a; ::); :; 3 2 1.0] 
 | u x y
-| -----
a| 3 2 1
b| 0 0 1

 Assign to columns: .[t;(::;`x`y); :; 9.0]
 | u x y
-| -----
a| 1 9 9
b| 0 9 9

2D Dictionary Do not Support Arithmetic, but keyed table support

-1"dd\na";
show a: `a`b!`u`x`y!/: 2 3 #til 6
-1"\nb";
show b: `b`c!`x`y`z!/:10* 2 3 #til 6
-1"\na+b";
show a+b
a
 | u x y
-| -----
a| 0 1 2
b| 3 4 5

b
 | x  y  z 
-| --------
b| 0  10 20
c| 30 40 50

a+b
'2021.08.15T22:05:11.029 mismatch
  [0]  \l /tmp/obq.q
       ^

2D Dictionary Matrix Multiplication

show x: `a`b!([] u: 2 1; x:0 2)
-1"\ny:";
show y: `u`x`y!([]c: 0 2 1; d: 1 0 1)
-1"\ncommonColumn";
show commonColumn: cols[value x] inter cols y
flip y[`u`x]
x[;`u`x]
flip[ y`u`x] {sum x*y}/:\: x[;`u`x]

.q.mm:{cc: cols[value x] inter cols y; flip[ y cc] {sum x*y}/:\: x[;cc]}
show x mm y
 | u x
-| ---
a| 2 0
b| 1 2

y:
 | c d
-| ---
u| 0 1
x| 2 0
y| 1 1

commonColumn
`u`x
c| 0 2
d| 1 0
a| 2 0
b| 1 2
 | a b
-| ---
c| 0 4
d| 2 1
 | a b
-| ---
c| 0 4
d| 2 1

Composite key D2D is supported by q , however, it can’t flip.

2D Dictionary with 2 key columns are supported, but can’t flip as expected.

show t: (`a`b;`c`d)! ([]x: 1 2; y:`c`d )
t[`a`b;]        /index supported
t[(`a;::);]     /but null doesn't mean *any* any more.
flip t
   | x y
---| ---
a b| 1 c
c d| 2 d
x| 1
y| `c
x| 0N
y| `
'2021.08.18T06:49:10.899 nyi
  [0]  flip t
       ^

D2D Usage Example: Edge Graph

Adjacency Matrix from Edge Matrix

D2D from csv

system "c 25 200"
cs: ("SSI",20#"B"; enlist",") 0: `:/Users/dh/d4m/examples/1Intro/2EdgeArt/EdgeUnix.csv
/ All column starts with V
vxx: {x where x like "V*"}cols cs
Nullify each `int$E: cs[`Edge]!flip vxx!cs vxx
  | V01 V02 V03 V04 V05 V06 V07 V08 V09 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19 V20
--| -------------------------------------------------------------------------------
B1| 1   1   1
S1| 1   1   1
G1| 1   1   1
O1| 1   1   1
O2| 1   1   1
P1| 1   1   1
B2|             1   1   1   1   1
S2|             1   1   1   1   1
G2|             1   1   1   1   1
O3|             1   1   1   1   1
O4|             1   1   1   1   1
P2|             1   1   1   1   1
O5|     1               1           1       1                   1               1
P3|                                     1   1       1   1
P4|     1                               1
P5|     1                                       1   1
P6|                                                         1   1   1
P7|                 1                                       1           1
P8|                         1                                   1           1

Edge to Edge. B1 can go to S1 by 3 nodes.

Nullify each E mm flip E
  | B1 S1 G1 O1 O2 P1 B2 S2 G2 O3 O4 P2 O5 P3 P4 P5 P6 P7 P8
--| --------------------------------------------------------
B1| 3  3  3  3  3  3                    1     1  1
S1| 3  3  3  3  3  3                    1     1  1
G1| 3  3  3  3  3  3                    1     1  1
O1| 3  3  3  3  3  3                    1     1  1
O2| 3  3  3  3  3  3                    1     1  1
P1| 3  3  3  3  3  3                    1     1  1
B2|                   5  5  5  5  5  5  1              1  1
S2|                   5  5  5  5  5  5  1              1  1
G2|                   5  5  5  5  5  5  1              1  1
O3|                   5  5  5  5  5  5  1              1  1
O4|                   5  5  5  5  5  5  1              1  1
P2|                   5  5  5  5  5  5  1              1  1
O5| 1  1  1  1  1  1  1  1  1  1  1  1  6  1  1  1  1     1
P3|                                     1  4  1  1
P4| 1  1  1  1  1  1                    1  1  2  1
P5| 1  1  1  1  1  1                    1  1  1  3
P6|                                     1           3  1  1
P7|                   1  1  1  1  1  1              1  3
P8|                   1  1  1  1  1  1  1           1     3

Node to Node. V01 can connect to V02 via 6 edges

Nullify each flip[E] mm E
   | V01 V02 V03 V04 V05 V06 V07 V08 V09 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19 V20
---| -------------------------------------------------------------------------------
V01| 6   6   6
V02| 6   9   6           1           1   1   1   1   1           1               1
V03| 6   6   6
V04|             6   6   6   6   6
V05|             6   7   6   6   6                           1           1
V06|     1       6   6   7   6   6   1       1                   1               1
V07|             6   6   6   7   6                               1           1
V08|             6   6   6   6   6
V09|     1               1           1       1                   1               1
V10|     1                               2   1       1   1
V11|     1               1           1   1   2       1   1       1               1
V12|     1                                       1   1
V13|     1                               1   1   1   2   1
V14|                                     1   1       1   1
V15|                 1                                       2   1   1   1
V16|     1               1   1       1       1               1   3   1       1   1
V17|                                                         1   1   1
V18|                 1                                       1           1
V19|                         1                                   1           1
V20|     1               1           1       1                   1               1

Select from cs where color = orange

convert from table to D2D
show cs1: cs[`Edge] ! (1#`Edge)_cs
  | Color  Order V01 V02 V03 V04 V05 V06 V07 V08 V09 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19 V20
--| --------------------------------------------------------------------------------------------
B1| Blue   2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
S1| Silver 2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
G1| Green  2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
O1| Orange 2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
O2| Orange 2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
P1| Pink   2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
B2| Blue   2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
S2| Silver 2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
G2| Green  2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
O3| Orange 2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
O4| Orange 2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
P2| Pink   2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
O5| Orange 1     0   1   0   0   0   1   0   0   1   0   1   0   0   0   0   1   0   0   0   1
P3| Pink   2     0   0   0   0   0   0   0   0   0   1   1   0   1   1   0   0   0   0   0   0
P4| Pink   2     0   1   0   0   0   0   0   0   0   1   0   0   0   0   0   0   0   0   0   0
P5| Pink   2     0   1   0   0   0   0   0   0   0   0   0   1   1   0   0   0   0   0   0   0
P6| Pink   2     0   0   0   0   0   0   0   0   0   0   0   0   0   0   1   1   1   0   0   0
P7| Pink   3     0   0   0   0   1   0   0   0   0   0   0   0   0   0   1   0   0   1   0   0
P8| Pink   3     0   0   0   0   0   0   1   0   0   0   0   0   0   0   0   1   0   0   1   0
select row whose color is Orange
#[;cs1] where  cs1[;`Color]=`Orange
  | Color  Order V01 V02 V03 V04 V05 V06 V07 V08 V09 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19 V20
--| --------------------------------------------------------------------------------------------
O1| Orange 2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
O2| Orange 2     1   1   1   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
O3| Orange 2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
O4| Orange 2     0   0   0   1   1   1   1   1   0   0   0   0   0   0   0   0   0   0   0   0
O5| Orange 1     0   1   0   0   0   1   0   0   1   0   1   0   0   0   0   1   0   0   0   1

D2D Examples: Entity Analysic

Load CSV

system "c 15 200"
ent: ("JS***"; enlist csv) 0: `:/Users/dh/d4m/examples/2Apps/1EntityAnalysis/Entity.csv
show ent: ((!) . 1#'`type`Type) xcol delete x from  update position: -1_'"J"$";" vs/:position from ent
doc                 entity        position   Type
-------------------------------------------------------
19960825_13108.txt  "addis ababa" 54 132 974 "LOCATION"
19960930_84704.txt  "addis ababa" ,60        "LOCATION"
19961004_96087.txt  "addis ababa" 61 305     "LOCATION"
19961006_98377.txt  "addis ababa" ,68        "LOCATION"
19961009_104796.txt "addis ababa" 59 443     "LOCATION"
19961010_107656.txt "addis ababa" ,61        "LOCATION"
19961031_158809.txt "addis ababa" ,2109      "LOCATION"
19961101_159647.txt "addis ababa" ,1485      "LOCATION"
19961113_185784.txt "addis ababa" ,62        "LOCATION"
19960821_6808.txt   "aden"        ,212       "LOCATION"
..

Explode table to sparse matrix, as matrix can’t handle string.

Combine 2 string columns to 1

show ent: delete entity, Type from update typeEnt: `$(Type,'"/",'entity) from ent
doc                 position   typeEnt
---------------------------------------------------
19960825_13108.txt  54 132 974 LOCATION/addis ababa
19960930_84704.txt  ,60        LOCATION/addis ababa
19961004_96087.txt  61 305     LOCATION/addis ababa
19961006_98377.txt  ,68        LOCATION/addis ababa
19961009_104796.txt 59 443     LOCATION/addis ababa
19961010_107656.txt ,61        LOCATION/addis ababa
19961031_158809.txt ,2109      LOCATION/addis ababa
19961101_159647.txt ,1485      LOCATION/addis ababa
19961113_185784.txt ,62        LOCATION/addis ababa
19960821_6808.txt   ,212       LOCATION/aden
..

Get all column name of the exploded matrix

show col: distinct asc ent`typeEnt
`s#`LOCATION/addis ababa`LOCATION/aden`LOCATION/adriatic sea`LOCATION/aegean sea`LOCATION/afghanistan`LOCATION/africa`LOCATION/akron`LOCATION/alabama`LOCATION/alaska`LOCATION/albania`LOCATION/alber..

Explode

ent1: ent[`doc]!exec col#/:(1#'typeEnt)!'(1#'position) from ent
system "c 20 200"
5#ent1
                   | LOCATION/addis ababa LOCATION/aden LOCATION/adriatic sea LOCATION/aegean sea LOCATION/afghanistan LOCATION/africa LOCATION/akron LOCATION/alabama LOCATION/alaska LOCATION/alban..
------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------..
19960825_13108.txt | 54                                                                                                                                                                              ..
19960930_84704.txt | 60                                                                                                                                                                              ..
19961004_96087.txt | 61                                                                                                                                                                              ..
19961006_98377.txt | 68                                                                                                                                                                              ..
19961009_104796.txt| 59                                                                                                                                                                              ..

What’s a column look like? ‘addis ababa’ is in 19960825_13108.txt at position 54

RmNull ent1[;`$"LOCATION/addis ababa"]
19960825_13108.txt | 54
19960930_84704.txt | 60
19961004_96087.txt | 61
19961006_98377.txt | 68
19961009_104796.txt| 59
19961010_107656.txt| 61
19961031_158809.txt| 2109
19961101_159647.txt| 1485
19961113_185784.txt| 62

D4M’s col2type

The Type column have different type: LOCATION, PERSION etc. convert each type to a column

The q way

distinct ent`Type
(,'/) 1!'value exec flip (`doc,first `$Type)!ent[i][`doc`entity] by `$Type from ent
"LOCATION"
"ORGANIZATION"
"PERSON"
"TIME"
doc                 location      
----------------------------------
19960825_13108.txt  "addis ababa" 
19960930_84704.txt  "addis ababa" 
19961004_96087.txt  "addis ababa" 
19961006_98377.txt  "addis ababa" 
19961009_104796.txt "addis ababa" 
19961010_107656.txt "addis ababa" 
19961031_158809.txt "addis ababa" 
19961101_159647.txt "addis ababa" 
19961113_185784.txt "addis ababa" 
19960821_6808.txt   "aden"        
19961026_145749.txt "aden"        
19961106_169278.txt "adriatic sea"
19961219_268288.txt "aegean sea"  
19960826_14771.txt  "afghanistan" 
19960910_44282.txt  "afghanistan" 
..
doc                | LOCATION       ORGANIZATION                                    PERSON                   TIME        
-------------------| ----------------------------------------------------------------------------------------------------
19960825_13108.txt | "addis ababa"  ""                                              ""                       "1996-08-25"
19960930_84704.txt | "addis ababa"  "united nations security council"               ""                       "1996-09-30"
19961004_96087.txt | "addis ababa"  "united nations high commissioner for refugees" ""                       "1996-10-04"
19961006_98377.txt | "addis ababa"  ""                                              ""                       "1996-10-06"
19961009_104796.txt| "addis ababa"  "united nations"                                ""                       "1996-10-09"
19961010_107656.txt| "addis ababa"  "united nations"                                "boutros boutros-ghali"  "1996-10-10"
19961031_158809.txt| "addis ababa"  "united nations"                                "andrew hill"            "1996-10-31"
19961101_159647.txt| "addis ababa"  "united nations"                                "salim ahmed salim"      "1996-11-01"
19961113_185784.txt| "addis ababa"  "united nations security council"               "salim ahmed salim"      "1996-11-13"
19960821_6808.txt  | "aden"         ""                                              ""                       "1996-08-21"
19961026_145749.txt| "aden"         ""                                              "abdul wali al-shumairi" "1996-10-26"
19961106_169278.txt| "adriatic sea" ""                                              ""                       "1996-11-06"
19961219_268288.txt| "aegean sea"   ""                                              ""                       "1996-12-19"
19960826_14771.txt | "afghanistan"  ""                                              "ahmad shah masood"      "1996-08-26"
19960910_44282.txt | "afghanistan"  ""                                              ""                       "1996-09-10"
..

col2type from Exploded 2D Dictionary

Where there is a value(position), spit the column name by /, and get a tuple. concatee the key with the tuple
key[ent1][0]
RmNull value[ent1][0]
key[ent1][0],/: "/" vs' string where not null value[ent1]0
`19960825_13108.txt
LOCATION/addis ababa| 54
`19960825_13108.txt "LOCATION" "addis ababa"
It takes 14 secconds to un explode a 0.1 billion row*col matrix. sparsity 0.02%
unExplode: { raze {x,/:"/"vs'string where not null y}'[key x; value x]}
ue: unExplode ent1
count[ue], distinct count each ue
ue
47089 3
`19960825_13108.txt  "LOCATION" "addis ababa"
`19960930_84704.txt  "LOCATION" "addis ababa"
`19961004_96087.txt  "LOCATION" "addis ababa"
`19961006_98377.txt  "LOCATION" "addis ababa"
`19961009_104796.txt "LOCATION" "addis ababa"
`19961010_107656.txt "LOCATION" "addis ababa"
`19961031_158809.txt "LOCATION" "addis ababa"
`19961101_159647.txt "LOCATION" "addis ababa"
`19961113_185784.txt "LOCATION" "addis ababa"
`19960821_6808.txt   "LOCATION" "aden"
`19961026_145749.txt "LOCATION" "aden"
`19961106_169278.txt "LOCATION" "adriatic sea"
`19961219_268288.txt "LOCATION" "aegean sea"
`19960826_14771.txt  "LOCATION" "afghanistan"
`19960910_44282.txt  "LOCATION" "afghanistan"
`19960910_44342.txt  "LOCATION" "afghanistan"
`19960912_49971.txt  "LOCATION" "afghanistan"
..
The exploded matrix has 47089 rows(.txt), 3657 columns, 0.17 billion cells, 47089 none null cell, sparsity: 0.027%
100* (0N!count ue) % 0N!(0N! count cols value ent1) * 0N!count ent1
Put each Type as a column for easy process.
Group by Type
first ue[;0 2] group ue[;1]
`19960825_13108.txt  "addis ababa"
`19960930_84704.txt  "addis ababa"
`19961004_96087.txt  "addis ababa"
`19961006_98377.txt  "addis ababa"
`19961009_104796.txt "addis ababa"
`19961010_107656.txt "addis ababa"
`19961031_158809.txt "addis ababa"
`19961101_159647.txt "addis ababa"
`19961113_185784.txt "addis ababa"
`19960821_6808.txt   "aden"
`19961026_145749.txt "aden"
`19961106_169278.txt "adriatic sea"
`19961219_268288.txt "aegean sea"
`19960826_14771.txt  "afghanistan"
`19960910_44282.txt  "afghanistan"
`19960910_44342.txt  "afghanistan"
`19960912_49971.txt  "afghanistan"
..
Convert from array of tuple to keyed table without lost information before uj
first { select `$entity by txt from flip `txt`entity !flip x}each  ue[;0 2] group ue[;1]
txt              | entity
-----------------| -------------------------------------------------------------------------------
19960820_2304.txt| `united states`washington
19960820_2324.txt| `britain`england`london
19960820_2344.txt| `britain`europe`germany`ireland`london
19960820_2374.txt| `london`new york
19960820_2414.txt| ,`brazil
19960820_2439.txt| `china`hong kong`indonesia`japan`korea`malaysia`singapore
19960820_2469.txt| `egypt`kuwait
19960820_2493.txt| `new york`united states
19960820_2515.txt| `argentina`brazil`detroit`new york`paraguay`south america`united states`uruguay
19960820_2539.txt| `san francisco`st. louis`united states`washington
19960820_2563.txt| `madrid`spain
19960820_2590.txt| `stockholm`sweden
19960820_2616.txt| ,`germany
19960820_2640.txt| `california`germany`japan`mexico`sweden`taiwan`united states
19960820_2659.txt| `atlanta`australia`italy
..
Rename entity column to value of Type
first {(enlist[`entity]!/:enlist'[`$key x]) xcol' value x} { select `$entity by txt from flip `txt`entity !flip x}each  ue[;0 2] group ue[;1]
txt              | LOCATION                                                                       
-----------------| -------------------------------------------------------------------------------
19960820_2304.txt| `united states`washington                                                      
19960820_2324.txt| `britain`england`london                                                        
19960820_2344.txt| `britain`europe`germany`ireland`london                                         
19960820_2374.txt| `london`new york                                                               
19960820_2414.txt| ,`brazil
19960820_2439.txt| `china`hong kong`indonesia`japan`korea`malaysia`singapore
19960820_2469.txt| `egypt`kuwait
19960820_2493.txt| `new york`united states
19960820_2515.txt| `argentina`brazil`detroit`new york`paraguay`south america`united states`uruguay
19960820_2539.txt| `san francisco`st. louis`united states`washington
19960820_2563.txt| `madrid`spain
19960820_2590.txt| `stockholm`sweden
19960820_2616.txt| ,`germany
19960820_2640.txt| `california`germany`japan`mexico`sweden`taiwan`united states
19960820_2659.txt| `atlanta`australia`italy
..

(uj/){(enlist[`entity]!/:enlist'[`$key x]) xcol' value x}  { select `$entity by txt from flip `txt`entity !flip x}each  ue[;0 2] group ue[;1]
txt              | LOCATION                                                                        ORGANIZATION PERSON                       TIME                  
-----------------| ------------------------------------------------------------------------------------------------------------------------------------------------
19960820_2304.txt| `united states`washington                                                                    ,`arshad mohammed            `1996-08-20`1997-09-01
19960820_2324.txt| `britain`england`london                                                                      `eddie george`kenneth clarke ,`1996-08-20
19960820_2344.txt| `britain`europe`germany`ireland`london                                                       `symbol$()                   `1996-06-30`1996-08-20
19960820_2374.txt| `london`new york                                                                             `symbol$()                   ,`1996-08-20
19960820_2414.txt| ,`brazil                                                                                     `symbol$()                   ,`1996-08-20
19960820_2439.txt| `china`hong kong`indonesia`japan`korea`malaysia`singapore                                    ,`lynne odonnell             ,`1996-08-20
19960820_2469.txt| `egypt`kuwait                                                                                `symbol$()                   ,`1996-08-20
19960820_2493.txt| `new york`united states                                                                      `mark wallace`mike spencer   ,`1996-08-20
19960820_2515.txt| `argentina`brazil`detroit`new york`paraguay`south america`united states`uruguay              ,`robert eaton               ,`1996-08-20
19960820_2539.txt| `san francisco`st. louis`united states`washington                                            ,`preston martin             ,`1996-08-20
19960820_2563.txt| `madrid`spain                                                                                `symbol$()                   ,`1996-08-20
19960820_2590.txt| `stockholm`sweden                                                                            `symbol$()                   ,`1996-08-20
19960820_2616.txt| ,`germany                                                                                    `symbol$()                   ,`1996-08-20
19960820_2640.txt| `california`germany`japan`mexico`sweden`taiwan`united states                                 ,`lisa raymond               ,`1996-08-20
19960820_2659.txt| `atlanta`australia`italy                                                                     ,`david prince               ,`1996-08-20
..