Add new fix method to set variable with element value #282

[TobiasNx]

Resolves #282

[blackwinter]

I've updated the implementation to account for missing values and added tests for edge cases (a52dc9b).

Open questions:

1. How to handle missing values? Is an empty string the right choice?
2. How to handle non-string values? Is the Java object representation the right choice? (Previously, you said that it "should only work with string-elements".)

Finally, the new function needs to be documented.

[TobiasNx]

I've updated the implementation to account for missing values and added tests for edge cases (a52dc9b).

Open questions:

1. How to handle missing values? Is an empty string the right choice?

I would not add a variable if the element does not exist.

2. How to handle non-string values? Is the Java object representation the right choice? (Previously, you [said](https://github.com/metafacture/metafacture-fix/issues/282#issuecomment-2355583039) that it "should only work with string-elements".)

I thought MF only processes string values in the Stream-Processing. Because of that we introduced marker for numbers in the json-encoder. The idea with string was the connection to the other ticket in mf-core but I have put the idea in core on hold.

Finally, the new function needs to be documented.

[blackwinter]

I thought MF only processes string values in the Stream-Processing.

Metafix knows data types strings, arrays and objects; variables can only contain strings.

[TobiasNx]

I thought MF only processes string values in the Stream-Processing.

Metafix knows data types strings, arrays and objects; variables can only contain strings.

Ah, I forgot. Then I would suggest an exception that the function expected a string but got hash or array.

[blackwinter]

I would not add a variable if the element does not exist.

This might become difficult to use:

1. If one record has the field and the next one doesn't, the variable keeps its previous value. We might want to remove the variable instead.
2. Referencing a non-existing variable (if we remove it or if it just hasn't been set yet) results in an error. So you can't reliably use the variable. We might have to introduce a conditional that explicitly checks for the presence of a variable; or add an option to specify the default value.

[TobiasNx]

I would not add a variable if the element does not exist.

This might become difficult to use:

1. If one record has the field and the next one doesn't, the variable keeps its previous value. We might want to remove the variable instead.

2. Referencing a non-existing variable (if we remove it or if it just hasn't been set yet) results in an error. So you can't reliably use the variable. We might have to introduce a conditional that explicitly checks for the presence of a variable; or add an option to specify the default value.

I favor the the option to define a default value. The conditional for the a variable sounds good. I am also in favor to delete the variable at the end of each record.

[blackwinter]

Then I would suggest an exception that the function expected a string but got hash or array.

Implemented in cee638a.

I favor the the option to define a default value.

Implemented in 46d1ef5.

The conditional for the a variable sounds good. I am also in favor to delete the variable at the end of each record.

Should probably be handled in separate issues.

[TobiasNx]

@TobiasNx I added documentation. Should @fsteeg do a code review?

[blackwinter]

Should @fsteeg do a code review?

Sure, why not.

[fsteeg]

Very useful feature!