Hi,

While invoking Spectrum Batch job using interface like JOBEXECUTOR.jar, we need to pass the Spectrum server username and password only in plain text format. Sample command are provided below :

While invoking Job in Dev Env:

java -jar jobexecutor.jar -j GCMR_US_Initial_MatchingEngine_Batch_Job -u ${USER} -p ${PASSWD} -h ${PB_SERVER} -s 45080 -w -d "$" -i 1 -t 9999 \
"Initial Matching Input"="file:${PB_PATH}/gcmr_${REGION}_${CMRCountry}_match_engine_inp_${FILE_DATE}.dat" \
"Initial Match Output file"="file:${PB_PATH}/gcmr_${REGION}_${CMRCountry}_match_engine_resp_${FILE_DATE}.dat"

While invoking Job in Prod Env over SSL connection:

java -Djavax.net.ssl.trustStore=C:\ak03148\spectrum\certs\spectrumKeyStore -jar jobexecutor.jar -h "vm-5a80-412a" -j ak03148_KANANAME -u admin -p admin -s 8443 "Read from File"="file:/tmp/gcmr_nap_jp_ecs_customer_pb_inp_name_20110906.dat" "Write to File"="file:/tmp/1.dat" -w -e

Now for both cases, we could only pass the plain text password with -p option. Now if we have to invoke the same in Prod Env, then anyone can list the job in unix and can view the password. This is not acceptable in secure environment and will have severe security audit related issues. We observed this problem and raised a PB ticket. It was told from PB Support that Spectrum version 7 or 8 could only handle password in plain text.
We are very near to Prod deployment of Spectrum dataflows for Citibank Japan, and need a quick product patch fix to handle this.

Please take this issue on priority.

When there is sorting on the columns names in the job, We see no results till the job is finished processing. This seems to be happening with the latest spectrum 8.0.

It will be good to see the results as the processing is happening as the sorting is on the reader.

It will be good, if "fields" in the "Read from file Options"  regerate based to column names rather than column numbers.

Hi

Request to tweak the behaviour of the following components.

ValidateAddress (for US)
1) Validate Address fails if the AddressLine and the other fields are not there as part of input.
2) If only Postal Code is given in input and is correct, the component should return correct city, stateprovince and county information.

ValidateAddressGlobal (for Global)
1) ValidateAddressGlobal should return the City, StateProvince and County corresponding to input Postal Code, if correct for all countries.

GeocodeUS
1) Presently Geocode Address for US returns county information if the input contains only PostalCode.
2) However it appends the word 'County' in the output.
3) This word county  is not seen appearing in the output from ValidateAddress County field.
4) Different components within PB should give consistent output.

GeocodeAddressGlobal
1) On inputting only PostalCode; for countries which have the County information, the tool is not giving.
2) Ideally on inputting PostalCode information which is correct, the behaviour should be similar to GeocodeUS, it should output the correct city, stateprovince and county.

Thank you

You should be able to override the connection/table information at runtime when using Read from DB and Write to DB stages in a job.  At the very least, allow the user to specify the connection/table name but ideally allow the user to also include schema file describing layout of the new source/sink tables.

Validate Address is currently configured to return the USPS Preferred city name.  Unfortunately, this does not always correspond to the USPS Primary city name for the given ZIP Code.  The user creates a hash of the address data with the Primary city name and stores it a database.  When identifying duplicate records, if the validated candidate has a USPS Preferred city that is different from the Primary city, it's not matched.

The customer would like an option in the Validate Address service to control whether the USPS Preferred or the USPS Default (Primary) city name is returned.

Ability to use CTRL+C & CTRL+V shortcuts and type in directly in the 'Schema file' and 'Data file' boxes of the Read From XML without having to click on the 'Open File' dialog. Same goes for Read from File, Write to File, Write to XML

Currenty the count of multiple matches is coming in two seperate variables in case of components ValidateAddress and ValidateAddressGlobal - the only difference being that in one case the variable has caps in one of the letters.

eg -
MultiMatchCount
MultimatchCount

Ideally only one of this should be having the value.

Several customers have requested the ability to export the Spectrum Execution History to a file (delimited, etc). This would enable a client to better assess usage and performance over time since a spreadsheet would provide calculative operations on values provided by Execution History as well as historical success/failure analysis.

Several customers have requested the ability to be able to execute a stored procedure from a dataflow.  It is currently possible to do this with Groovy/JDBC code in a Transformer stage, but having this ability in a stage would simplify dataflow development.

Enable the FME Workbench to be OEM'd as a module to Stratus would make end-to-end workflows more integrated in the one platform.  At the moment a customer would need to make a call out of Spectrum to something like FME Server and back in - making workflows disjointed and unable to be visualised end-to-end in the one designer window.

Business case for the customer is the performance and results from using a platform with best fo breed modules.  Also improved time to value (integrated workflows), TCO (one architecture), and flexibility (access to a range of specialist spatial transformers). 

Potential for customers to import base workflows authored in FME and augment in Spectrum with transformers only available in Spectrum.  If can import FME workflows from ESRI platforms this would decrease migration risk and cost.

ESRI OEM the FME visual workflow for the data format transformers (but exclude the FME raster processing and use their own).
PBS already OEM FME in Mapinfo Professional as part of the universal translator but do not enable the visual workflow (which may be reasonable given the need to make it easy to use).

A large ServiceStar customer is interested in a configuration setting in Spectrum that limits the number of connections accepted: for example when a threshold of 200,000 connections is reached any new connections are rejected.

Usability Enhancement: To be able to rename a Dataflow Job or Service by double-clicking the (<New Dataflow> (Job)) tab in the Enterprise Designer. Same as you can rename a spreadsheet from within Microsoft Excel

Read from XML, Read from File, Write to XML, etc. should support UNC path
 
\\host\dir\file.txt

The original user uncleansed record is send to PB for standardization. PB generates multiple matches.

Now the user selects one of the records and can modify the record or send the same again for restandardizing.

In the present scenario, if the user sends same record for standardizing though the field which has resulted in multiple matches is fed into address line input of PB, the tool is again generating multiple matches.

Ideally if the information which has resulted in multiple matches (eg- locality) is fed into input of PB, it should only output one address.

Currenlty only high and low ranges for units (secondary info) is available as a Qualifier for output data when a unit is present on input. 
The customer is converting from EnvinsaOnline to Spectrum service and would like to be able to know if the address that is sent on input has valid APT information.  They would like to see HighUnitNumber and LowUnitNumber populated for those records where a possible secondary information exists.

Curently after standardization,we get abbreviated state/province values in the output for Validate Address(US) component.

There should be an option to set Output return values for State/province as 'Extended' similar to the case of ValidateAddressGlobal.

EX:

Currently ValidateAddress returns :NJ
Expected output from component is :New Jersey.

I would like to be able to pass in a start and finish point with a bunch of unordered points in between (each with a unique id) and get back an ordered list of points for an optimised route (with the unique id attached) so I can see how my points were re-ordered.

When calling Spectrum to "Get Travel Directions" with the "optmise intermediate points" options turned on, the route engine returns a route that optimises the intermediate points by re-ordering them.  However the information returned doesn't actually show how the points were re-ordered (ie what the new order is).

When running a FindNearest in Spectrum using a list of customer's points and a MapInfo .tab file containing polygons, the service behaves in the following way:

-For each point, it finds the distance to the edge of the nearest polygon in the boundary file specified.
-For the nearest polygon, the "Point-of-Interest" geometry returned by FindNearest represents the centroid of the polygon, and not the edge itself.

If FindNearest can find the nearest edge and calculate the distance to this point, it should also return the geometry at the edge of the polygon and not the centroid, or we should at least have the option of selecting which geometry we want returned.