How to implement a PATCH with `database/sql`? - sql

Let’s say you have a basic API (GET/POST/PATCH/DELETE) backed by an SQL database.
The PATCH call should only update the fields in the JSON payload that the user sends, without touching any of the other fields.
Imagine the table (let's call it sample) has id, string_a and string_b columns, and the struct which corresponds to it looks like:
type Sample struct {
ID int `json:"id"`
StringA string `json:"stringA"`
StringB string `json:"stringB"`
}
Let's say the user passes in { "stringA": "patched value" } as payload. The json will be unmarshalled to something that looks like:
&Sample{
ID: 0,
StringA: "patched value",
StringB: "",
}
For a project using database/sql, you’d write the query to patch the row something like:
// `id` is from the URL params
query := `UPDATE sample SET string_a=$1, string_b=$2 WHERE id=$3`
row := db.QueryRow(query, sample.StringA, sample.StringB, id)
...
That query would update the string_a column as expected, but it’d also update the string_b column to "", which is undesired behavior in this case. In essence, I’ve just created a PUT instead of a PATCH.
My immediate thought was - OK, that’s fine, let’s use strings.Builder to build out the query and only add a SET statement for those that have a non-nil/empty value.
However, in that case, if a user wanted to make string_a empty, how would they accomplish that?
Eg. the user makes a PATCH call with { "stringA": "" } as payload. That would get unmarshalled to something like:
&Sample{
ID: 0,
StringA: "",
StringB: "",
}
The “query builder” I was theorizing about would look at that and say “ok, those are all nil/empty values, don’t add them to the query” and no columns would be updated, which again, is undesired behavior.
I’m not sure how to write my API and the SQL queries it runs in a way that satisfies both cases. Any thoughts?

I think reasonable solution for smaller queries is to build UPDATE query and list of bound parameters dynamically while processing payload with logic that recognizes what was updated and what was left empty.
From my own experience this is clear and readable (if repetitive you can always iterate over struct members that share same logic or employ reflection and look at struct tags hints, etc.). Every (my) attempt to write universal solution for this ended up as very convoluted overkill supporting all sorts of corner-cases and behavioral differences between endpoints.
func patchSample(s Sample) {
var query strings.Builder
params := make([]interface{}, 0, 2)
// TODO Check if patch makes sense (e.g. id is non-zero, at least one patched value provided, etc.
query.WriteString("UPDATE sample SET")
if s.StringA != "" {
query.WriteString(" stringA = ?")
params = append(params, s.StringA)
}
if s.StringB != "" {
query.WriteString(" stringB = ?")
params = append(params, s.StringB)
}
query.WriteString(" WHERE id = ?")
params = append(params, s.ID)
fmt.Println(query.String(), params)
//_, err := db.Exec(query.String(), params...)
}
func main() {
patchSample(Sample{1, "Foo", ""})
patchSample(Sample{2, "", "Bar"})
patchSample(Sample{3, "Foo", "Bar"})
}
EDIT: In case "" is valid value for patching then it needs to be distinguishable from the default empty value. One way how to solve that for string is to use pointer which will default to nil if value is not present in payload:
type Sample struct {
ID int `json:"id"`
StringA *string `json:"stringA"`
StringB *string `json:"stringB"`
}
and then modify condition(s) to check if field was sent like this:
if s.StringA != nil {
query.WriteString(" stringA = ?")
params = append(params, *s.StringA)
}
See full example in playground: https://go.dev/play/p/RI7OsNEYrk6

For what it's worth, I solved the issue by:
Converting the request payload to a generic map[string]interface{}.
Implementing a query builder that loops through the map's keys to create a query.
Part of the reason I went this route is it fit all my requirements, and I didn't particularly like having *strings or *ints laying around.
Here is what the query builder looks like:
func patchQueryBuilder(id string, patch map[string]interface{}) (string, []interface{}, error) {
var query strings.Builder
params := make([]interface{}, 0)
query.WriteString("UPDATE some_table SET")
for k, v := range patch {
switch k {
case "someString":
if someString, ok := v.(string); ok {
query.WriteString(fmt.Sprintf(" some_string=$%d,", len(params)+1))
params = append(params, someString)
} else {
return "", []interface{}{}, fmt.Errorf("could not process some_string")
}
case "someBool":
if someBool, ok := v.(bool); ok {
query.WriteString(fmt.Sprintf(" some_bool=$%d,", len(params)+1))
params = append(params, someBool)
} else {
return "", []interface{}{}, fmt.Errorf("could not process some_bool")
}
}
}
if len(params) > 0 {
// Remove trailing comma to avoid syntax errors
queryString := fmt.Sprintf("%s WHERE id=$%d RETURNING *", strings.TrimSuffix(query.String(), ","), len(params)+1)
params = append(params, id)
return queryString, params, nil
} else {
return "", []interface{}{}, nil
}
}
Note that I'm using PostgreSQL, so I needed to provide numbered parameters to the query, eg $1, which is what params is used for. It's also returned from the function so that it can be used as follows:
// Build the patch query based on the payload
query, params, err := patchQueryBuilder(id, patch)
if err != nil {
return nil, err
}
// Use the query/params and get output
row := tx.QueryRowContext(ctx, query, params...)

Related

How do perform a graph query and join?

I apologize for the title, I don't exactly know how to word it. But essentially, this is a graph-type query but I know RavenDB's graph functionality will be going away so this probably needs to be solved with Javascript.
Here is the scenario:
I have a bunch of documents of different types, call them A, B, C, D. Each of these particular types of documents have some common properties. The one that I'm interested in right now is "Owner". The owner field is an ID which points to one of two other document types; it can be a Group or a User.
The Group document has a 'Members' field which contains an ID which either points to a User or another Group. Something like this
It's worth noting that the documents in play have custom IDs that begin with their entity type. For example Users and Groups begin with user: and group: respectively. Example IDs look like this: user:john#castleblack.com or group:the-nights-watch. This comes into play later.
What I want to be able to do is the following type of query:
"Given that I have either a group id or a user id, return all documents of type a, b, or c where the group/user id is equal to or is a descendant of the document's owner."
In other words, I need to be able to return all documents that are owned by a particular user or group either explicitly or implicitly through a hierarchy.
I've considered solving this a couple different ways with no luck. Here are the two approaches I've tried:
Using a function within a query
With Dejan's help in an email thread, I was able to devise a function that would walk it's way down the ownership graph. What this attempted to do was build a flat array of IDs which represented explicit and implicit owners (i.e. root + descendants):
declare function hierarchy(doc, owners){
owners = owners || [];
while(doc != null) {
let ownerId = id(doc)
if(ownerId.startsWith('user:')) {
owners.push(ownerId);
} else if(ownerId.startsWith('group:')) {
owners.push(ownerId);
doc.Members.forEach(m => {
let owner = load(m, 'Users') || load(m, 'Groups');
owners = hierarchy(owner, owners);
});
}
}
return owners;
}
I had two issues with this. 1. I don't actually know how to use this in a query lol. I tried to use it as part of the where clause but apparently that's not allowed:
from #all_docs as d
where hierarchy(d) = 'group:my-group-d'
// error: method hierarchy not allowed
Or if I tried anything in the select statement, I got an error that I have exceeded the number of allowed statements.
As a custom index
I tried the same idea through a custom index. Essentially, I tried to create an index that would produce an array of IDs using roughly the same function above, so that I could just query where my id was in that array
map('#all_docs', function(doc) {
function hierarchy(n, graph) {
while(n != null) {
let ownerId = id(n);
if(ownerId.startsWith('user:')) {
graph.push(ownerId);
return graph;
} else if(ownerId.startsWith('group:')){
graph.push(ownerId);
n.Members.forEach(g => {
let owner = load(g, 'Groups') || load(g, 'Users');
hierarchy(owner, graph);
});
return graph;
}
}
}
function distinct(value, index, self){ return self.indexOf(value) === index; }
let ownerGraph = []
if(doc.Owner) {
let owner = load(doc.Owner, 'Groups') || load(doc.Owner, 'Users');
ownerGraph = hierarchy(owner, ownerGraph).filter(distinct);
}
return { Owners: ownerGraph };
})
// error: recursion is not allowed by the javascript host
The problem with this is that I'm getting an error that recursion is not allowed.
So I'm stumped now. Am I going about this wrong? I feel like this could be a subquery of sorts or a filter by function, but I'm not sure how to do that either. Am I going to have to do this in two separate queries (i.e. two round-trips), one to get the IDs and the other to get the docs?
Update 1
I've revised my attempt at the index to the following and I'm not getting the recursion error anymore, but assuming my queries are correct, it's not returning anything
// Entity/ByOwnerGraph
map('#all_docs', function(doc) {
function walkGraph(ownerId) {
let owners = []
let idsToProcess = [ownerId]
while(idsToProcess.length > 0) {
let current = idsToProcess.shift();
if(current.startsWith('user:')){
owners.push(current);
} else if(current.startsWith('group:')) {
owners.push(current);
let group = load(current, 'Groups')
if(!group) { continue; }
idsToProcess.concat(group.Members)
}
}
return owners;
}
let owners = [];
if(doc.Owner) {
owners.concat(walkGraph(doc.Owner))
}
return { Owners: owners };
})
// query (no results)
from index Entity/ByOwnerGraph as x
where x.Owners = "group:my-group-id"
// alternate query (no results)
from index Entity/ByOwnerGraph as x
where x.Owners ALL IN ("group:my-group-id")
I still can't use this approach in a query either as I get the same error that there are too many statements.

Do strings need to be escaped inside parametrized queries?

I'm discovering Express by creating a simple CRUD without ORM.
Issue is, I'm not able to find any record through the Model.findBy() function
model User {
static async findBy(payload) {
try {
let attr = Object.keys(payload)[0]
let value = Object.values(payload)[0]
let user = await pool.query(
`SELECT * from users WHERE $1::text = $2::text LIMIT 1;`,
[attr, value]
);
return user.rows; // empty :-(
} catch (err) {
throw err
}
}
}
User.findBy({ email: 'foo#bar.baz' }).then(console.log);
User.findBy({ name: 'Foo' }).then(console.log);
I've no issue using psql if I surround $2::text by single quote ' like:
SELECT * FROM users WHERE email = 'foo#bar.baz' LIMIT 1;
Though that's not possible inside parametrized queries. I've tried stuff like '($2::text)' (and escaped variations), but that looks far from what the documentation recommends.
I must be missing something. Is the emptiness of user.rows related to the way I fetch attr & value ? Or maybe, is some kind of escape required when passing string parameters ?
"Answer":
As stated in the comment section, issue isn't related to string escape, but to dynamic column names.
Column names are not identifiers, and therefore cannot be dynamically set using a query parameter.
See: https://stackoverflow.com/a/50813577/11509906

Boolean fields not getting updated

I'm trying to update my entire database by batching all the changes in my request using Updates from gorm. All the fields are getting updated except for the bool values. No error is being thrown too by the database.
This is my struct
type Client struct {
ClientID string `json:"client_id"`
FirstName string `json:"first_name"`
LastName string `json:"last_name"`
Email string `json:"email"`
Telephone string `json:"telephone"`
AcceptNotification bool `json:"accept_notification"`
Verfied bool `json:"verified"`
}
and the code to update client is
func (b *Client) UpdateClient(request models.Client) (err error) {
db := b.DB
result := db.Table("client").
Where("client_id = ? AND store_id = ?", request.ClientID, request.StoreID).
Updates(request)
if result.Error != nil {
return result.Error
}
return nil
}
In the documentation it says
When update with struct, GORM will only update non-zero fields, you might want to use map to update attributes or use Select to specify fields to update
What could be the issue? What is the solution?
You can't update zero value using struct (you should use a map instead). From the doc:
Updates multiple columns Updates supports update with struct or
map[string]interface{}, when updating with struct it will only update
non-zero fields by default
// Update attributes with `struct`, will only update non-zero fields
db.Model(&user).Updates(User{Name: "hello", Age: 18, Active: false})
// UPDATE users SET name='hello', age=18, updated_at = '2013-11-17 21:34:10' WHERE id = 111;
// Update attributes with `map`
db.Model(&user).Updates(map[string]interface{}{"name": "hello", "age": 18, "active": false})
// UPDATE users SET name='hello', age=18, active=false, updated_at='2013-11-17 21:34:10' WHERE id=111;
NOTE When update with struct, GORM will only update non-zero fields,
you might want to use map to update attributes or use Select to
specify fields to update

Gorm add multiple slices in inserting in a many to many

I'm new using go and gorm. I'm trying to insert many values in one SQL query.
I wrote this query to add multiple conversations to a user:
relationUserConversation := make([][]uint, len(users))
for i, v := range users {
relationUserConversation[i] = []uint{conversation.ID, v}
}
result = r.db.Debug().Exec(
"INSERT INTO `user_has_conversations` (`user_has_conversations`.`conversation_id`, `user_has_conversations`.`user_id`) VALUES ?",
relationUserConversation, // If i do this it works relationUserConversation[0], relationUserConversation[1]
// The issue is because the query has this value "VALUES ((35,1),(35,2))", but should be to work (35,1),(35,2)
)
I also tried to add it directly with the conversation that would be what I would like to do, but I'm having issue trying to add the relation with the many to many because instead of creating the relation between the user and the conversation it tries to add the user.
My conversation model:
type Conversation struct {
ID uint `gorm:"primarykey"`
Users []*User `gorm:"many2many:user_has_conversations;"`
Messages []ConversationMessage
}
Would be great if i could create a new conversation with the related users in one query instead of creating first the conversation and after the relation to the users.
Below is a minimum working example using the Gorm Appends method (see documentation here) to create a many to many association between two (or more) models. Hopefully you can adapt this to your use case.
package main
import (
"fmt"
"gorm.io/driver/sqlite"
"gorm.io/gorm"
)
type User struct {
gorm.Model
Name string
Conversations []Conversation `gorm:"many2many:user_conversations;"`
}
type Conversation struct {
gorm.Model
Name string
Users []*User `gorm:"many2many:user_conversations;"`
}
func main() {
db, err := gorm.Open(sqlite.Open("many2many.db"), &gorm.Config{})
if err != nil {
panic("failed to connect database")
}
// Migrate the schema
err = db.AutoMigrate(&User{}, &Conversation{})
if err != nil {
fmt.Print(err)
}
userOne := User{
Name: "User One",
}
userTwo := User{
Name: "User Two",
}
// Create users
db.Create(&userOne)
db.Create(&userTwo)
conversation := Conversation{
Name: "Conversation One",
}
// Create conversation
db.Create(&conversation)
// Append users
err = db.Model(&conversation).Association("Users").Append([]User{userOne, userTwo})
if err != nil {
fmt.Print(err)
}
for _, convUser := range conversation.Users {
fmt.Println("Hello I am in the conversation: " + convUser.Name)
}
// Clean up database
db.Delete(&userOne)
db.Delete(&userTwo)
db.Delete(&conversation)
}
Number of queries
If you enable Debug() on Gorm:
err = db.Debug().Model(&conversation).Association("Users").Append([]User{userOne, userTwo})
It shows this:
[0.144ms] [rows:2] INSERT INTO `user_conversations`
(`conversation_id`,`user_id`) VALUES (8,15),(8,16) ON CONFLICT DO NOTHING
The Values part is correct (what you were trying to do manually) and achieved using the ORM.

Why does this documentation example fail? Is my workaround an acceptable equivalent?

While exploring the documented example raised in this perl6 question that was asked here recently, I found that the final implementation option - (my interpretation of the example is that it provides three different ways to do something) - doesn't work. Running this;
class HTTP::Header does Associative {
has %!fields handles <iterator list kv keys values>;
sub normalize-key ($key) { $key.subst(/\w+/, *.tc, :g) }
method EXISTS-KEY ($key) { %!fields{normalize-key $key}:exists }
method DELETE-KEY ($key) { %!fields{normalize-key $key}:delete }
method push (*#_) { %!fields.push: #_ }
multi method AT-KEY (::?CLASS:D: $key) is rw {
my $element := %!fields{normalize-key $key};
Proxy.new(
FETCH => method () { $element },
STORE => method ($value) {
$element = do given $value».split(/',' \s+/).flat {
when 1 { .[0] } # a single value is stored as a string
default { .Array } # multiple values are stored as an array
}
}
);
}
}
my $header = HTTP::Header.new;
say $header.WHAT; #-> (Header)
$header<Accept> = "text/plain";
$header{'Accept-' X~ <Charset Encoding Language>} = <utf-8 gzip en>;
$header.push('Accept-Language' => "fr"); # like .push on a Hash
say $header<Accept-Language>.perl; #-> $["en", "fr"]
... produces the expected output. Note that the third last line with the X meta-operator assigns a literal list (built with angle brackets) to a hash slice (given a flexible definition of "hash"). My understanding is this results in three seperate calls to method AT-KEY each with a single string argument (apart from self) and therefore does not exersise the default clause of the given statement. Is that correct?
When I invent a use case that excersises that part of the code, it appears to fail;
... as above ...
$header<Accept> = "text/plain";
$header{'Accept-' X~ <Charset Encoding Language>} = <utf-8 gzip en>;
$header{'Accept-Language'} = "en, fr, cz";
say $header<Accept-Language>.perl; #-> ["en", "fr", "cz"] ??
# outputs
(Header)
This Seq has already been iterated, and its values consumed
(you might solve this by adding .cache on usages of the Seq, or
by assigning the Seq into an array)
in block at ./hhorig.pl line 20
in method <anon> at ./hhorig.pl line 18
in block <unit> at ./hhorig.pl line 32
The error message provides an awesome explanation - the topic is a sequence produced by the split and is now spent and hence can't be referenced in the when and/or default clauses.
Have I correctly "lifted" and implemented the example? Is my invented use case of several language codes in the one string wrong or is the example code wrong/out-of-date? I say out-of-date as my recollection is that Seq's came along pretty late in the perl6 development process - so perhaps, this code used to work but doesn't now. Can anyone clarify/confirm?
Finally, taking the error message into account, the following code appears to solve the problem;
... as above ...
STORE => method ($value) {
my #values = $value».split(/',' \s+/) ;
$element = do given #values.flat {
when 1 { $value } # a single value is stored as a string
default { #values } # multiple values are stored as an array
}
}
... but is it an exact equivalent?
That code works now (Rakudo 2018.04) and prints
$["en", "fr", "cz"]
as intended. It was probably a bug which was eventually solved.