Error Message Guide

UCO/Lick Mask System

The problem with error messages is that they are usually unhelpful. Too terse, too packed with cryptic numbers. The programmer has desperately attempted to squeeze as much message at all into a small bottle, to the detriment of comprehensibility. This guide should help.

Error 1 in HDU H (N): TFIELDS F does not match actual # of cols A

A FITS Table Extension, of which there are several in your mask file, is described by headers in its HDU. The TFIELDS keyword tells how many fields are in each row of the table. This message indicates that the number of columns actually found does not match the value in the TFIELDS keyword.

What this means is that your FITS file is internally inconsistent. In the HDU whose name is N and which is number H (from 1 to some small integer number) in the file, the value F of TFIELDS does not match the actual value A of columns found in the table.

This almost certainly indicates a serious bug in program which wraps up your mask design into the FITS file. You should submit a tar file containing all the mask design subfiles as well as the broken FITS file to sla@ucolick.org.

Error 2 in HDU H (N): NAXIS2 R does not match actual # of rows A

The NAXIS2 HDU keyword for a FITS table extension tells how many rows are expected in the table. This message indicates that a number of rows was found different from the value of NAXIS2 for that table extension.

What this means is that your FITS file is internally inconsistent. In the HDU whose name is N and which is number H (from 1 to some small integer number) in the file, the value R of NAXIS2 does not match that actual value A of rows encountered.

Error 3 in HDU H (N): Col I (CN) type is CT, but TFORMi is TF

Every FITS table extension is self-documenting. Each column in the table has a format which "belongs" to it, but is also documented in the HDU by means of the keyword TFORMi. If the two disagree, the file is internally inconsistent.

In HDU number H, named N, column number I (named CN), has a type of CT, but the corresponding value of TFORMi does not agree.

This almost certainly indicates a serious bug in the program which wraps up your mask design into the FITS file. You should submit a tar file containing all the mask design subfiles as well as the broken FITS file to sla@ucolick.org.

Error 4 in HDU H (N): NULL value for col I (CN) at row R not permitted

A special keyword TNULLi is used in the header of a FITS table extension, to specify whether null values are permitted in the corresponding table column. If TNULL1 is 0, then nulls are not permitted in column 1.

This message means that in HDU number H, named N, there is a null value for column I (named CN) at row R in the data. This is illegal because TNULLi is zero.

While this might indicate an error in the grouper code that wraps your design files into a single FITS file, it's more likely that you edited a null or blank entry into the data at some point in the proceedings. You should check your source files carefully, and remember whether you might have used fv to edit the FITS file (we don't recommend this).

If you can't figure out how that null got into your data, submit all the data files and the fits file to sla@ucolick.org.

Error 5 in HDU H (N): Col C CN is CT but row R value is V

The self-documenting nature of a FITS table extension includes strict "typing" of all columns. If a column is designated as an integer, then no strings or floating-point values must appear in it.

This message indicates that in HDU number H, named N, column C named CN is of type CT, but in row R the value of column C is V, which doesn't match type CT. In other words, you have a string where you should have a number, or a float where you should have an int.

You should check your data files to make sure they contain what you think they do, and particularly suspect any manual edits you may have attempted on the FITS file itself.

Error 6 foreign key FC (col C in T1), table T2 does not exist

When we need to express relationships of structure, hierarchy, or "belonging" in databases and FITS files, we use relational concepts like primary key and foreign key. Quite simply, a primary key is the simplest attribute necessary to determine uniqueness. For sake of argument we could say that for people this is the social security number. If a person has filed 10 tax forms over the last 5 years, each of those tax forms has that person's number on it. The tax forms (if you put all their data into a table) would have a Foreign Key (social security number) which was the Primary Key in another table (People).

There are many relationships of this kind in the mask data. Each Slit is associated with one or more objects. Each Design is associated with some number of Slits. And so on. What this error indicates is that in Table T1, column C has been designated as a Foreign Key. We expect it to exist as column FC in the other table T2. But the other table T2 doesn't exist! Or in other words, we have a tax form, but we've lost the filing cabinet with all the taxpayer personal data in it, including their social security numbers.

This is a serious error and indicates major failure in the mask design software or the wrapper program. You should mail this message (as many of them as you see in the log) to sla@ucolick.org, phillips@ucolick.org, and de@ucolick.org.

Error 7 col FC does not exist in table T2 (foreign key XREF for col C of T1)

(See explanation of foreign keys above, in Error 6)

This message indicates that although table T1 has a column C which is designated as a foreign key, we cannot find the corresponding column FC in table T2. In other words, the tax form has a social security number on it, but social security number is not one of the attributes we know about for taxpayers.

Error 8 foreign key value V in col C rec R of table T1 not found in fkey table T2 for col FC

(See explanation of foreign keys above, in Error 6)

This message indicates that although table T1 has a column C which is designated as a foreign key, and table T2 exists and has a column FC which is its primary key, we've found a value of T1 column C at row R which doesn't correspond to any value of FC in T2. In other words, we have a tax form here with a social security number on it that doesn't match any of the numbers of the taxpayers in our files.

Warning 9 in HDU H (N): col C (CN, CT) but row R value is V

While superficially similar to Error 5, this is merely a warning. It warns that the data in column C, named CN, of type CT, at row R, seems a little odd for its datatype. For example, you specified a floating point value but the actual value is an integer. This is not a fatal error, because FITS' format scan mechanism is based on FORTRAN and is incredibly forgiving. However, it's an oddity that you might want to look at more closely.

de@ucolick.org
De Clarke