distinct sql and ignore field in filters and sorting

[miklemv]

Now by default, the sql request with the distinct. I think this is not very good, better distict = false.
I added argument distict to request and set default distict = false. The default mode can be changed static setter. I'm not sure that distinct is needed in arguments.
Example request "{ Books(distinct: true) { select { genre } }}" return 5 record. This is misleading.

Added @GraphQLIgnoreFilter - prohibiting the use of field in filters.
Added @GraphQLIgnoreOrder - prohibiting the use of the field for sorting.
Using all fields in filters and sorting is dangerous, because this can create an unexpected strong load on the database if the field does not have an index. It seems to me that it is better to list the fields explicitly for sorting and filtering, but this will change the current behavior.

[codecov]

Codecov Report

Merging #95 into master will increase coverage by 0.24%.
The diff coverage is 71.42%.

@@             Coverage Diff              @@
##             master      #95      +/-   ##
============================================
+ Coverage     65.44%   65.68%   +0.24%     
- Complexity      349      366      +17     
============================================
  Files            37       37              
  Lines          1962     2011      +49     
  Branches        291      296       +5     
============================================
+ Hits           1284     1321      +37     
- Misses          552      562      +10     
- Partials        126      128       +2

Impacted Files	Coverage Δ	Complexity Δ
...ventures/graphql/jpa/query/example/books/Book.java	0% <ø> (ø)	0 <0> (ø)	⬇️
.../query/example/books/BooksSchemaConfiguration.java	100% <100%> (ø)	4 <0> (ø)	⬇️
.../query/schema/impl/GraphQLJpaQueryDataFetcher.java	93.25% <62.5%> (-1.87%)	21 <1> (+1)
...query/autoconfigure/GraphQLJpaQueryProperties.java	80% <66.66%> (-5.72%)	9 <2> (+2)
...jpa/query/schema/impl/GraphQLJpaSchemaBuilder.java	86.43% <70.27%> (-1.32%)	103 <10> (+12)
...a/query/schema/impl/QraphQLJpaBaseDataFetcher.java	72.05% <0%> (+0.88%)	111% <0%> (+2%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4c8bf31...ac2ef24. Read the comment docs.

[igdianov]

Let's configure default distinct with GraphJPASchemaBuilder method using Spring Boot configuration properties in starter, so folks can customize it via application properties or Java configuration.

[igdianov]

@miklemv There is a rather big change in #96 that will cause conflicts with this PR. Please, wait making any changes until I merge it into master tomorrow. Then, you can rebase your PR to resolve conflicts.

[miklemv]

path: graphql - will clean
graphql-jpa-query-example-merge/pom.xml - dublicate dependency )

I create options in GraphQLJpaQueryProperties and GraphQLSchemaBuilder config methods.

[miklemv]

rebase branch and fix ignore field in sub filterers.

[igdianov]

@miklemv Looks very good, but I think it makes more sense to put distinct argument inside select field. It is more logical place. I also wonder if the distinct argument will support variable bindings?

[miklemv]

@igdianov
I think that this parameter is not needed at the graphQL schea level. The
{ Books(distinct: true) { select { genre } }}
will not return 2 and 5 records, since the id field is involved in the sql query. Function SetDistinctFetcher at the developer level is enough.
Now it's difficult to make a distinct for the graphQL fields and I think this is not very useful functionality.

[igdianov]

@miklemv Sounds good! We can always refactor later if needed. Thank you!

[chanhengseang3]

I think distinct parameter does not work correctly:

Is there any known issues about this?

[chanhengseang3]

[miklemv]

@chanhengseang3
Yes. I I talked about it
#95 (comment)

Thish problem is hibernate level. I don'k know how distinc operation with models by fields.