Date Documentation
- I need to document different types of dates.
Overview
Documenting the when of data is often more complicated than the date the resource was created or modified. It is also important to know when the data collection took place and how long the project lasted. Dates are contained in different dialects on different ways.
Implementation
Content Standard for Digital Geospatial Metadata (CSDGM)
Structure
Publish Date <pre<noinclude></noinclude>><citation>
<citeinfo>
<origin>/metadata/idinfo/citation/citeinfo/origin</origin>
<pubdate>/metadata/idinfo/citation/citeinfo/pubdate</pubdate>
<pubtime>/metadata/idinfo/citation/citeinfo/pubtime</pubtime>
<title>/metadata/idinfo/citation/citeinfo/title</title>
<edition>/metadata/idinfo/citation/citeinfo/edition</edition>
<geoform>/metadata/idinfo/citation/citeinfo/geoform</geoform>
<serinfo>
<sername>/metadata/idinfo/citation/citeinfo/serinfo/sername</sername>
<issue>/metadata/idinfo/citation/citeinfo/serinfo/issue</issue>
</serinfo>
<pubinfo>
<pubplace>/metadata/idinfo/citation/citeinfo/pubinfo/pubplace</pubplace>
<publish>/metadata/idinfo/citation/citeinfo/pubinfo/publish</publish>
</pubinfo>
<othercit>/metadata/idinfo/citation/citeinfo/othercit</othercit>
<onlink>/metadata/idinfo/citation/citeinfo/onlink</onlink>
<lworkcit>
<citeinfo/>
</lworkcit>
</citeinfo>
</citation>
Range of Dates <pre<noinclude></noinclude>> <timeinfo>
<rngdates>
<begdate>20000101</begdate>
<enddate>20000202</enddate>
</rngdates>
</timeinfo>
Single Date <pre<noinclude>> <timeinfo>
<sngdate>
<caldate>44440404</caldate>
</sngdate>
</timeinfo>
Usage
| Usage | Description and Xpath |
|---|---|
| Date of publication for resource being documented | /metadata/idinfo/citation/citeinfo/pubdate |
| Range of dates during which the data was collected | /metadata/idinfo/timeperd/timeinfo/rngdates/begdate And /metadata/idinfo/timeperd/timeinfo/rngdates/enddate |
| Single date during which the data was collected | /metadata/idinfo/timeperd/timeinfo/sngdate/enddate |
The time information dates can be used to describe the dataset in idinfo/timeperd, the lineage of the dataset in dataqual/lineage and cross references in idinfo/crossref The publish date in the citation can occur wherever citations can occur in a CSDGM record.
Directory Interchange Format (DIF)
Structure
Creation / Revision Date <pre<noinclude></noinclude>><Data_Set_Citation>
<Dataset_Creator>/DIF/Data_Set_Citation/Dataset_Creator</Dataset_Creator> <Dataset_Editor>/DIF/Data_Set_Citation/Dataset_Editor</Dataset_Editor> <Dataset_Title>/DIF/Data_Set_Citation/Dataset_Title</Dataset_Title> <Dataset_Series_Name>/DIF/Data_Set_Citation/Dataset_Series_Name</Dataset_Series_Name> <Dataset_Release_Date>/DIF/Data_Set_Citation/Dataset_Release_Date</Dataset_Release_Date> <Dataset_Release_Place>/DIF/Data_Set_Citation/Dataset_Release_Place</Dataset_Release_Place> <Dataset_Publisher>/DIF/Data_Set_Citation/Dataset_Publisher</Dataset_Publisher> <Version>/DIF/Data_Set_Citation/Version</Version> <Issue_Identification>/DIF/Data_Set_Citation/Issue_Identification</Issue_Identification> <Data_Presentation_Form>/DIF/Data_Set_Citation/Data_Presentation_Form</Data_Presentation_Form> <Other_Citation_Details>/DIF/Data_Set_Citation/Other_Citation_Details</Other_Citation_Details> <Dataset_DOI>/DIF/Data_Set_Citation/Dataset_DOI</Dataset_DOI> <Online_Resource>/DIF/Data_Set_Citation/Online_Resource</Online_Resource>
</Data_Set_Citation>
Dataset Range of Dates <pre<noinclude></noinclude>><Temporal_Coverage>
<Start_Date>/DIF/Temporal_Coverage/Start_Date</Start_Date> <Stop_Date>/DIF/Temporal_Coverage/Stop_Date</Stop_Date>
</Temporal_Coverage>
Paleo Dates <pre<noinclude></noinclude>><Paleo_Temporal_Coverage>
<Paleo_Start_Date>/DIF/Paleo_Temporal_Coverage/Paleo_Start_Date</Paleo_Start_Date>
<Paleo_Stop_Date>/DIF/Paleo_Temporal_Coverage/Paleo_Stop_Date</Paleo_Stop_Date>
<Chronostratigraphic_Unit uuid="uuid11">
<Eon>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Eon</Eon>
<Era>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Era</Era>
<Period>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Period</Period>
<Epoch>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Epoch</Epoch>
<Stage>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Stage</Stage>
<Detailed_Classification>/DIF/Paleo_Temporal_Coverage/Chronostratigraphic_Unit/Detailed_Classification</Detailed_Classification>
</Chronostratigraphic_Unit>
</Paleo_Temporal_Coverage>
Usage
| Usage | Description and Xpath |
|---|---|
| Date of publication for resource being documented | /DIF/Data_Set_Citation/Dataset_Release_Date |
| Range of dates during which the data was collected | /DIF/Temporal_Coverage/Start_Date And /DIF/Temporal_Coverage/Stop_Date |
| Paleo Dates | /DIF/Paleo_Temporal_Coverage/Paleo_Start_Date And /DIF/Paleo_Temporal_Coverage/Paleo_Stop_Date |
EOS Clearinghouse (ECHO)
Structure
Usage
International Organization for Standardization 19115 (ISO19115)
There are many dates included in the ISO Metadata Standards and they have several different types - each with its own characteristics. This page has information about valid formats for those dates. Ron Lake's blog includes a helpful [description] of time in GML.
Structure
gco:DateTime <pre<noinclude>> <gmd:dateTime>
<gco:DateTime>2001-01-01T00:00:00</gco:DateTime>
</gmd:dateTime>
gmd:CI_Date <pre<noinclude>> <gmd:CI_Date>
<gmd:date>
<gco:Date>2000-01-01</gco:Date>
</gmd:date>
<gmd:dateType>
<gmd:CI_DateTypeCode
codeList="http://www.isotc211.org/2005/resources/Codelist/gmxCodelists.xml#CI_DateTypeCode"
codeListValue="creation">creation</gmd:CI_DateTypeCode>
</gmd:dateType>
</gmd:CI_Date>
On-Going Datasets <pre<noinclude>> <gmd:date>
<gmd:CI_Date>
<gmd:date gco:nilReason="inapplicable"/>
<gmd:dateType>
<gmd:CI_DateTypeCode codeList="http://www.isotc211.org/2005/resources/Codelist/gmxCodelists.xml#CI_DateTypeCode"
codeListValue="publication">publication</gmd:CI_DateTypeCode>
</gmd:dateType>
</gmd:CI_Date>
</gmd:date> ... <gmd:status>
<gmd:MD_ProgressCode codeListValue="onGoing" codeList="http://www.isotc211.org/2005/resources /codeList.xml#MD_ProgressCode">onGoing</gmd:MD_ProgressCode>
</gmd:status> ... <gmd:temporalElement>
<gmd:EX_TemporalExtent id="boundingTemporalExtent">
<gmd:extent>
<gml:TimePeriod gml:id="d1e50">
<gml:beginPosition>2003-01-18T00:23:00.000Z</gml:beginPosition>
<gml:endPosition indeterminatePosition="now"/>
</gml:TimePeriod>
</gmd:extent>
</gmd:EX_TemporalExtent>
</gmd:temporalElement>
Relative Times <pre<noinclude>> <gmd:temporalElement>
<gmd:EX_TemporalExtent>
<gmd:extent>
<gml:TimePeriod gml:id="id">
<gml:duration>P1D</gml:duration>
<gml:endPosition indeterminatePosition="now"/>
</gml:TimePeriod>
</gmd:extent>
</gmd:EX_TemporalExtent>
</gmd:temporalElement>
Usage
| Usage | Description and Xpath |
|---|---|
| Citation for resource being documented | /metadata/idinfo/citation/citeinfo |
| Citation for cross reference. | /metadata/idinfo/crossref/citeinfo |
| Citation for source used in lineage. | /metadata/dataqual/lineage/srcinfo/srccite/citeinfo |
gco:Date Type
Date: gives values for year, month and day. Character encoding of a date is a string which shall follow the format for date specified by [ISO 8601]. A full date is formatted as YYYYMMDD or YYYY-MM-DD. This type is used in the following fields:
- MI_Metadata.dateStamp - The ISO definition of this field is "date that the metadata was created". Common usage seems to be evolving towards "date that the metadata was created or updated" because it seems reasonable that, once the metadata have been updated, the date of that update is more interesting than the original creation date. Also, the update actually creates the current metadata, so this usage seems consistent with the ISO definition. Note: ISO 19115-1 replaces the datestamp/Date field with the dateinfo/CI_Date object, which enables the type of date to be specified. ISO 19115-1 expands the date type codelist from 3 values to 16.
- MD_MaintenanceInformation.dateOfNextUpdate
- CI_Citation.editionDate - The ISO edition field is synonymous with the concept of version, so this is the release date for the version of the resource being cited.
gco:DateTime Type
DateTime: combination of a date and a time type (given by an hour, minute and second). Character encoding of a DateTime shall follow [ISO 8601]. Combined dates and times should be formatted as YYYYMMDDThh:mm:ss, YYYYMMDDThhmmss, or YYYY-MM-DDThh:mm:ss. These representations include no TimeZone indicator, so they are assumed to be local time. YYYYMMDDThh:[mm:ssZ] would indicate universal time.
- MD_Usage.usageDateTime
- DQ_Element.dateTime
- LI_ProcessStep.dateTime
- MD_StandardOrderProcess.plannedAvailableDateTime
Note: This element was incorrectly defined in the ISO 19139 nschema as an [xs:dateTime]. That type does not allow all of the ISO 8601 options. Specifically, it does not allow the specification of a time range. It will likely be deprecated in the revision of the standard and replaced with stepDateTime.
gmd:CI_Date Type
This type is used only in the CI_Citation and is the only date type that includes a code from the CI_DateTypeCode codelist. Valid values from the CI_DateTypeCode CodeList are: creation (001), publication (002), and revision (003). Note: ISO 19115-1 adds an additional 13 date type code list values to the code list. In figure, see values highlighted in red.
~~ATTACHED_IMAGE~~DateTypeCodes.PNG On-Going Datasets
Many datasets are collected continuously through time. In these cases, a clear publication date does not exist. This situation can be indicated by a combination of "inapplicable" for the citation date, a status code = onGoing, and an endPosition with indeterminatePosition=now:
Relative Times
In some situations it is necessary to express a time in terms of a fixed period before the present. For example, on-going data processing systems might use observation sets and products from the previous day in the creation of a product for today. In these cases the duration and the end of the time period are known, so the XML looks like:
See [ISO 8601] for information on durations.
Uncertain Times
The GML elements for describing time positions (beginPosition and endPosition) include the indeterminatePosition attribute that can have values of before, after, now, and unknown. This attribute can be used with a time value to express uncertain times. For example, a dataset that began sometime before 1980 and continues to the present could be described as above.
=Notes
The FGDC Standard allows a variety of date representations including "Free Date" in many locations. A number of assumptions must be made in order to translate these dates into the more structured formats required by ISO. If using XSLT's to translate FGDC date content to ISO (such as the NOAA transforms), logic must be added to achieve the highest level of interoperability and and highest rate of valid ISO input. These assumptions might be based on the length of the FGDC date string:
4 characters = YYYY 6 characters = YYYYMM or bcYYYY 7 characters = 'unknown' 8 characters = YYYYMMDD or bcYYYYMM 10+ characters = bcYYYYMMDD or ccYYYYYYY... or cdYYYYYYY...
This DECODE query is used in selecting dates dates with YYYYMMDD, YYYYMM and YYYY formats from Oracle (other formats are invalid):
decode(LENGTH(DATETIME),6,'<gml:Date>'||SUBSTR(DATETIME,1,4)||'-'||SUBSTR(DATETIME,5,2)||'</gml:Date>', 8,'<gml:Date>'||DATETIME||'</gml:Date>',4,'<gml:Date>'||DATETIME||'</gml:Date>', '<gml:Date>"'||DATETIME||'" is an invalid date format</gml:Date>')DATETIME
Other assumptions must be made as well such as
- accounting for non-standard dates - accounting for various capilatizations such as 'Unknown' or 'unknown' or 'UNKNOWN' - accounting for common misspellings such as 'unkown', 'unknow' ,'unkwon', etc. - accounting for extra spaces - accounting for input dates that start with bc, cc, or cd
The indeterminatePosition can be used in a couple of cases if the date is unknown or unpublished
<gml:timePosition indeterminatePosition="unknown"> <gml:timePosition indeterminatePosition="now">
If the date can not be recognized, set it as a reason for having no date:
<gmd:dateTime gco:nilReason="Date Content"/>.
Of course, many of these will not be valid values for nilReason so they will be identified as problems during record validation. For the NOAA transforms, if a date is missing or cannot not be determined, gco:nilReason will be set to 'unknown' by default. Ending dates of 'present' and 'unknown' (also taking into consideration capitalizations, misspelling, and extra spaces) will use indeterminatePosition.
