Upload log messages

This is a listing of messages you may encounter in the Review the Upload log. Messages that are classified as ERROR may lead to the failure of creating a properly functioning Spark service.

Excel file messages

These atypical settings on an Excel file will automatically stop the compilation process as the user would see different results compared to Spark.

• ERROR “Use 1904 date system” is incompatible with Spark. This setting can be disabled in Excel for Windows under “Advanced Options” or Excel for Mac in “Preferences”.

• ERROR “Set precision as displayed” is incompatible with Spark. This setting can be disabled in Excel for Windows under “Advanced Options” or Excel for Mac in “Preferences”.

X mapping messages

During the upload process Spark also reviews the X mappings to ensure they will not lead to errors in the Spark service.

• ERROR The following named ranges include cells containing content that is more than 32,767 characters in length and may not be processed correctly. Please reduce the number of characters in the identified cells.

• Spark has identified a potential spelling error in the listed named ranges.

• Multiple definitions have been defined for these inputs. Please rename these inputs so that they have only a single definition.

• Multiple definitions have been defined for these outputs. Please rename these outputs such that they have only a single definition.

• The following named ranges contain Excel references that could not be found (likely due to a deleted sheet or cell(s)) and will not be processed correctly. Please resolve the Excel errors.

• The following named ranges contain Excel errors and will not be processed correctly. Please resolve the Excel errors.

• The following named ranges exist outside of the worksheet’s used range. Spark will not process all of the upload warnings for these ranges.

• The following named ranges contain Data Validation that resolve to Excel errors and will not be processed correctly.

• The following Xcall references a service that does not exist in this tenant.

• The Xrichoutput definition contains an incorrect instruction which will not output correctly.

Xinput messages

• ERROR The following named ranges contain duplicated headers and cannot be converted. Please ensure that every column header is populated and unique.

• The following input ranges overlap and may lead to incorrect calculations.

• The following named ranges contain a dynamic array headers and may not provide a correct response.

• The following input named ranges contain formulas. The formulas will be evaluated and the calculated values will be used as default values for these inputs.

• The following input named input ranges include 2000 rows of entries. Spark will not process all of the upload warnings for this range.

Xsolve messages

• The following solve parameters contain Excel errors and are preventing the solve from being executed correctly. Please resolve the Excel errors.

• The Xoutput has been defined as the “by changing” value in an Xsolve and will prevent the solve from being executed correctly. Please adjust this output to refer to another cell.

• The following solve parameter is missing and prevent the solve from being executed correctly. Please resolve the Excel errors.

Neuron compiler messages

During the upload process Neuron, our "Excel-to-code" compiler may generate a number of messages. Below is a guide to the messages you may encounter.

Self-test

INFO [Summary] Spark performs a self-test comparing the saved output values in the Excel file and a calculation of the formulas using Spark. The self-test can be used to understand and diagnose mismatches. For the mapped outputs in this workbook, there were 1 matches out of 1 (100.0%) assessed at 12 digits accuracy. The maximum difference has a significance of -∞ and magnitude of -∞. Self-test results may not match if the workbook includes external references, random functions, Xcall, Xrichoutput, Xsolve. The outputs that contain a mismatch are listed in the compiler messages. Read the Spark Neuron documentation for more information about the self-test.

When an Excel file is saved, the result of the calculation in each cell is saved into the Excel file. Upon the completion of the Neuron compilation, a self-test will be performed on Spark-mapped output cells where:

1. Neuron will perform a calculation using the values saved into the Spark-mapped inputs.

2. The Neuron calculated results are taken from the Spark mapped output cells (value).

3. Excel's calculated value for the Spark mapped output cells are defined as (expected).

4. Spark compares expected and value for each mapped output cell.

   • If expected = value then this output cell counts as a match.

   • If expected <> value then:

     • The magnitude of this difference is determined as magnitude = log10(abs(expected - value)).

     • The significance of this difference is determined as significance = log10(abs(expected - value)) - log10(abs(expected)).

       • If significance < -1 * (accuracy check significant digits), then the output cell would also count as a match.

5. The summary of the self-test is written in inspected output cells and max diff significance, and max diff magnitude.

INFO [Summary] Spark performs a self-test comparing the saved output values in the Excel file and a calculation of the formulas using Spark. The self-test can be used to understand and diagnose mismatches. Because this service includes the use of Xcall, the self-test was not performed. We recommend additional testing of this Spark service against the Excel calculations. Read the Spark Neuron documentation for more information about the self-test.

If an Excel file contains Using CS.SPARK_XCALL() UDF functions, the self-test will be skipped.

Othe reasons the self-test may not come back as 100% may be due to:

• Random functions or volatile time functions such as NOW() and TODAY().

• If there are no mapped outputs in the file.

External links

INFO Worksheet Sheet1: contains 1 external reference which was replaced with value(s) evaluated by that reference.

• Cells containing simple external workbook links will be converted to a value, e.g. ='[Source File.xlsx]FX Rates'!$E$6.

• However, cells containing formulas that use external workbook links in an Excel function are likely to be evaluated as #REF!.

Unsupported functions

When a function is unsupported in Neuron, you may see an error like this when you compile:

ERROR Cell Sheet1!K22: contains an unrecognized function: BahtText. Consider an alternative function.

Sometimes the compiler may have an issue with a formula and raise compiler messages containing a formula name. For example in the example below, the GETPIVOTDATA function is preventing the compilation from occurring.

"wasm-ld.bin: error: CMakeFiles/checkfile.dir/checkfile.cpp.o: undefined symbol: Workbook::GetPivotData(CV const&, Address3d const&, std::initializer_list) const"

• User-defined functions from Visual Basic for Applications or other Excel add-ins will not be recognized. The cells containing these functions will be evaluated as #NAME?.

• Spark does provide extensive support for Excel's LAMBDA functions which enable users to create their own user-defined functions.

• Please refer to Neuron function support for our development roadmap.

Formula messages

During the compilation process, Neuron will perform a deeper check on the formulas used in the workbook and may raise additional messages.

• ERROR Cell and adjacent cells use the CELL() function but one parameter is missing. Populate both parameters of the CELL() function to use it with Spark.

  ERROR Cell and adjacent cells contain an unrecognized function. Consider an alternative function.

• ERROR Xsolve references a constant value in the change by parameter. This prevents the Xsolve from being calculated. Please replace this with formula referencing a single cell.

• Cell and adjacent cells contain a Named Range referring to another worksheet that could not be accessed. Error will be used as the computed value for these cells.

• Cell and adjacent cells contain references to an invalid named range. Error will be used as the computed value for these cells.

• Cell and adjacent cells contain references to an invalid named range: Error will be used as the computed value for these cells.

• Cell and adjacent cells contain the INFO() function, which is not supported by Spark. Outputs that rely upon this formula may not be correct.

• Cell and adjacent cells contain an unrecognized function. Error will be used as the computed value for this function. Consider using an alternative function.

• Cell and adjacent cells contain an unrecognized function: Error will be used as the computed value for the function. Consider using an alternative function.

• Cell and adjacent cells include the use of INDIRECT() in Excel, which may lead to lower performance in Spark. Consider alternative formulas such as CHOOSE(), MATCH(), INDEX(), OFFSET().

• Worksheet contains Data Validation that is beyond the used range of the worksheet. Data Validation for these cells will not be processed.

• Worksheet contains Data Validation that is not mapped as a Spark input. Data Validation for these cells will not be processed.

• Worksheet contains external reference(s) which were replaced with the value in the saved Excel file.

• Worksheet uses "Lotus compatibility Settings Transition formula evaluation", which is incompatible with Spark. This can be disabled in Excel's Advanced Options.

• Xsolve contains an invalid header. Please check the Xsolve block to ensure the proper settings are used.