A property helps to configure a document meta. It defines how to process it and store
it as a dedicated index field. It is a functional layer over the technical configuration.
For example, if you want to search for the content of a specific meta, the property is going
to store the meta as a dedicated index field and configure the search to enable you to do so.
Properties in a data model class have a Data type, such as
alphanumeric, numeric, or geographic. The type, as well as other settings, determines what
kind of elements you automatically create from the property when you scan your data sources.
Semantic types are assigned to all alphanumeric properties in a data model class, to
determine the type of semantic processing applied to these properties at index time.
You can create the following
types of properties:
Alphanumeric
Numerical: integer, unsigned integer, and double
Date: date only and date-time
Geographic: GPS point and XY point
Measure
Indexing Options for All Properties
When creating a data model property, you can choose one of the following field types:
This section explains the options available depending on the property type. The following
schema represents it.
Indexing Options for Data Model Properties
Note:
For an example explaining what is exactly created in the index schema and search logic,
see What is the Data Model Expansion.
Indexing Options for Index Fields
The most common indexing method is to assign a dedicated index field to the property. The
property can then be:
Searchable which means that user queries can be applied to this
field. If the property is searchable with a prefix:
If this class is the default class, you can search for the property with a prefix
format of property:value.
Otherwise, you can search for the property with a prefix format of
classname_property:value.
Retrievable, which means that the field can display in the
search results. A hit meta with the property name and value appears in hit content.
Indexing Option for Facets
The property can also be stored as a Category facet.
A category is created with the value of the property.
Faceted navigation is automatically created for this property. In that case, it is
possible to search for the value of the property.
If the property belongs to the default class, you can search for the property
with a prefix format of property:value.
Otherwise, you can search for the property with a prefix format of
classname_property:value.
Note:
If the property is both in a facet and in a dedicated field, search is handled
through the field. The facet is then only used for faceted navigation.
Indexing Options for Dedicated Fields and Facets
The property can be made searchable without prefix, which means in
addition to the indexing determined by other options, the property is indexed as
searchable in the field called text.
This allows users to search this property without specifying a prefix in the query.
When this option is selected, the Relevance class select box
displays, to let you adjust the ranking level of the index field. The value can be a
standard ranking key from -1 to 8 (8 being the highest rank) or a custom value with a
positive integer higher than 8.
If the property is not in a dedicated index field or in a facet, or searchable in text,
it can still be stored and retrieved, but not searched for.
Advanced Options
All properties have advanced parameters. For more details, see AdvancedParams in the Exalead CloudView XML Configuration
Reference Guide.
Indexing Options for Alphanumeric Properties
When
adding a new alphanumeric property, you must assign it a semantic
type. It defines the global semantic processing to be applied when indexing
document metas. You can then customize this semantic processing at the Analysis pipeline
level (Index > Data Processing > Semantic Processors).
Exalead CloudView includes default semantic types, but you can modify them or create additional ones in
Data Model > Semantic Types.
Table 1. Default Semantic Types
Semantic type
Description
text
Apply this semantic type to document metas containing several words or sentences.
For example, you can apply this semantic type for a meta with a value like:
“Deep Blue was a chess-playing computer. It is known for being the first
piece of artificial intelligence to win both a chess game and a chess match
against a reigning world champion under regular time controls.”
You must do it if you want to perform extra processing treatment in the analysis
pipeline.
metadata
Apply this semantic type to document metas containing a few words that do not
require language detection and spell-checking.
For example, if we have the meta value Deep Blue
Searching for deep blue in lowercase would work, as well as
searching for deep only or for blue only.
identifier
Apply this semantic type to document metas containing IDs.
With this semantic type, document metas are searched in their normalized forms,
but they cannot be tokenized and the language cannot be detected.
For example, if we have the meta value Deep Blue
Searching for “deep blue” (with quotes) would work, but
searching for deep only or blue only would not
work.
url
Apply this semantic type to document metas containing web addresses.
Indexing Options for Numerical Properties
When storing a numerical property as a facet, only equality search is possible. Range
search is only possible in a dedicated index field. For more information, see Numerical Range facets.
When storing a numerical property as a dedicated index field, searchable with
prefix includes searching using the range operators.
Indexing Options for Date Properties
When adding a date property, you must define the input format that
triggers the indexing of dates. You can leave the field empty for an automatic detection of
standard formats or specify the date format to detect in the document corpus.
Exalead CloudView uses the UNIX date syntax to specify date and time formats for date fields in the data
model. The same syntax is also used in the search logic for date metas displayed in the hit
content, and dynamic date facets.
Note:
There is also a time output format which can be set for each meta
to display in the search results hits. Its default value is %m/%d/%Y
%H:%M:%S. See Search Logics > Hit Content > Metas
and expand, for example, the lastmodifieddate meta Time
format operation.
Note:
If a timezone is detected inside the date time value, it allows you to convert and store
this value in UTC format. For more information, see Specifying a Timezone for Date Time Metas.
Table 2. UNIX Date Syntax
Specifier
Description
Values/Example
Day
%a
weekday, abbreviated
Mon
%A
weekday, full
Monday
%d
day of the month (dd), zero filled
08
%e
day of the month (dd)
8
%j
day of year, zero filled
001-366
%u
day of week starting with Monday (1), that is, mtwtfss
7 (for Sunday)
%w
day of week starting with Sunday (0), that is, smtwtfs
0 (for Sunday)
Week
%U
week number Sunday as first day of week
00–53
%W
week number Monday as first day of week
01–53
%V
week of the year
01–53
Month
%m
mm month
08
%h
Mon
Aug
%b
Mon, locale's abbreviated
Aug
%B
locale's full month, variable length
August
Year
%y
yy two-digit year
00–99
%Y
ccyy year
2014
%g
2-digit year corresponding to the %V week number
%G
4-digit year corresponding to the %V week number
Century
%C
cc century
00–99
Date
%D
mm/dd/yy
08/20/14
%x
locale's date representation (mm/dd/yy)
08/20/2014
%F
%Y-%m-%d
2014-08-20
Hours
%l
hour (12 hour)
5
%I
hour (12 hour) zero filled
05
%k
hour (24 hour)
17
%H
hour (24 hour) zero padded
17
%p
locale's capitalize AM or PM (blank in many locales)
PM
%P
locale's lowercase am or pm
pm
Minutes
%M
MM minutes
18
Seconds
%s
seconds since 00:00:00 1970-01-01 UTC (UNIX epoch)
1345483096
%S
SS second
00–60 (The 60 is required to accommodate a leap second)
Time
%r
hours, minutes, seconds (12-hour clock)
05:18:16 PM
%R
hours, minutes (24-hour clock)
17:18
%T
hours, minutes, seconds (24-hour clock)
17:18:16
%X
locale's time representation
11:07:26 AM
Date and Time
%c
locale's date and time
Sat Nov 04 12:02:33 EST 1989
Indexing Options for Geographical Properties
Exalead CloudView supports two types of geographic coordinates:
GPS fields, also called WGS84, store pairs of latitudes and
longitudes separated by commas. They are expressed in decimal format, with an accuracy
of 6 decimal places. Example: 37.818667,-122.478383 is a valid
meta for GPS fields.
XY fields, also called Meter, store pairs of integers separated
by commas. As no unit is defined, you can consider the unit as meters, miles, or
whatever unit you need. Example: 125, 8215 is a valid meta for XY
fields.
You can select Use separate metas for each coordinate if latitude/ X
and Longitude/ Y come from two different metas in your data source. Exalead CloudView concatenates the two coordinates to store them in a single index field.
Indexing Options for Measure Properties
To handle a meta-data that contains a measure, add a property in your data model with a
Measure data type. Then specify the unit
symbol of your measure; this works as indexing unit and also as default input
unit.
By default, the expansion control configuration automatically specifies all required
elements for the good indexing & search of this new property:
a Units of Measurement Normalizer document processor in the
Data Processing > MODEL_NAME > Document Processors.
a measurement prefix handler with the name of your measure property in
Search Logics > SEARCH_LOGIC > Query Language.
