Keyword Matching
1001
the quik brwn fox jumps over the lzy dog
1002
jaded zombies actted quaintly - but kept driving ther oxen frward
1003
crasy fredrick bouht many vry7 exquisite opal jewles .
1004
we promptly judged antique ivery bucles for the next priz.
1005
jinxed wizards pluk , ivy from the big quilt6
Human entered strings
1
the
2
quick
3
brown
4
fox
5
jumps
6
over
7
lazy
8
dog
9
jaded
10
zombies
11
acted
12
quaintly
13
but
14
kept
15
driving
16
their
17
oxen
18
forward
19
crazy
20
frederick
21
bought
22
many
23
very
24
exquisite
25
opal
26
jewels
27
we
28
promptly
29
judged
30
antique
31
ivory
32
buckles
33
for
34
next
35
prize
36
jinxed
37
wizards
38
pluck
39
ivy
40
from
41
big
42
quilt
Dictionary keywords
String
Parse into words
Custom
Warning
All
ManualDelayed
Union human entered words to dictionary lookup
RecordID
String
RecordID
Fuzzy match words - tweak dataset depending on entry values
Join dictionary words back to fuzzy match result
Only bring in unique words to keep from redundant fuzzy match results
Unique: Word
Join back to original strings so that words have lookup ID after fuzzy match
UniqueValue
Tile on unique value to get word placement in sentence structure
Unique Value
Sort by string id and tile sequence num to recombine strings in proper order
ID - Ascending
Tile_SequenceNum - Ascending
Keep only unique matches
Unique: RecordID, RecordID2
Remove unwanted fields
Warning
All
ByName
Includes un-fuzzy-matched words (if there are any) from input to help re-configure fuzzy match
Recombine strings from fuzzy match suggestions
C:\ProgramData\Alteryx\Engine\Engine_2476_766e60c45ae04bc5ba667a14e638c618_\Engine_4528_346b1454f49c4bb991284a54545d765b_.yxdb
String
Replace
String
Warn
String_Matched
Only keep unique words
Unique: String
length(trim([String])) >= 2
Custom
Filter out puntuation that isn't contained in word strings
length(trim([String])) >= 2
2000
RecordID
Int32
0
Give words ID rejoin for recombination with fuzzy match words later
length([String]) >= 2
Custom
Filter out puntuation that isn't contained in word strings
length([String]) >= 2
2000
RecordID
Int32
0
Give words record id for fuzzy match and recombination later
Only keep unique words
Unique: String
Fuzzy match dictionary words vs human entered
Determine original string order
Use dictionary of all correct keywords to match human entered strings to
Clean human entered strings before fuzzy match if possible - can provide flexibility in fuzzy match configuration
Make sure string values used in human entered data match post-parsed (workflow arm below) record ids exactly for matching later
Horizontal
Fuzzy Keyword Match