I have ran a crawler on json S3 file for updating an existing external table.
Once finished I checked the SVL_S3LOG to see the structure of the external table and saw it was updated and I have new column with Array<int> type like expected.
When I have tried to execute select * on the external table I got this error: "Invalid operation: Nested tables do not support '*' in the SELECT clause.;"
So I have tried to detailed the select statement with all columns names:
select name, date, books.... (books is the Array<int> type)
from external_table_a1
and got this error:
Invalid operation: column "books" does not exist in external_table_a1;"
I have also checked under "AWS Glue" the table external_table_a1 and saw that column "books" is recognized and have the type Array<int>.
Can someone explain why my simple query is wrong?
What am I missing?
Querying JSON data is a bit of a hassle with Redshift: when parsing is enabled (eg using the appropriate SerDe configuration) the JSON is stored as a SUPER type. In your case that's the Array<int>.
The AWS documentation on Querying semistructured data seems pretty straightforward, mentioning that PartiQL uses "dotted notation and array subscript for path navigation when accessing nested data". This doesn't work for me, although I don't find any reasons in their SUPER Limitations Documentation.
Solution 1
What I have to do is set the flags set json_serialization_enable to true; and set json_serialization_parse_nested_strings to true; which will parse the SUPER type as JSON (ie back to JSON). I can then use JSON-functions to query the data. Unnesting data gets even crazier because you can only use the unnest syntax select item from table as t, t.items as item on SUPER types. I genuinely don't think that this is the supposed way to query and unnest SUPER objects but that's the only approach that worked for me.
They described that in some older "Amazon Redshift Developer Guide".
Solution 2
When you are writing your query or creating a query Redshift will try to fit the output into one of the basic column data types. If the result of your query does not match any of those types, Redshift will not process the query. Hence, in order to convert a SUPER to a compatible type you will have to unnest it (using the rather peculiar Redshift unnest syntax).
For me, this works in certain cases but I'm not always able to properly index arrays, not can I access the array index (using my_table.array_column as array_entry at array_index syntax).
Related
I'm building a framework for rust-postgres.
I need to know what value type will be returned from a row.try_get, to get the value in a variable of the appropriate type.
I can get the sql type from row.columns()[index].type, but not if the value is nullable , so i can't decide to put the value in a normal type or a Option<T>.
I can use just the content of the row to understand it, i can't do things like "get the table structure from Postgresql".
is there a way?
The reason that the Column type does not expose any way to find out if a result column is nullable is because the database does not return this information.
Remember that result columns are derived from running a query, and that query may contain arbitrary expressions. If the query was a simple SELECT of columns from a table, then it would be reasonably simple to determine if a column could be nullable.
But it could also be a very complex expression, derived from multiple columns, subselects or even custom functions. Postgres can figure out the data type of each column, but in the general case it doesn't know if a result column may contain nulls.
If your application is only performing simple queries, and you know which table column each result column comes from, then you can find out if that table column is nullable like this:
SELECT is_nullable
FROM information_schema.columns
WHERE table_schema='myschema'
AND table_name='mytable'
AND column_name='mycolumn';
If your queries are not that simple then I recommend you always get the result as an Option<T> and handle the possibility that the result might be None.
I have a basic REST service backed by a PostgreSQL database with a table with various columns, one of which is a JSONB column that contains arbitrary data. Clients can store data filling in the fixed columns and provide any JSON as opaque data that is stored in the JSONB column.
I want to allow the client to query the database with constraints on both the fixed columns and the JSONB. It is easy to translate some query parameters like ?field=value and convert that into a parameterized SQL query for the fixed columns, but I want to add an arbitrary JSONB query to the SQL as well.
This JSONB query string could contain SQL injection, how can I prevent this? I think that because the structure of the JSONB data is arbitrary I can't use a parameterized query for this purpose. All the documentation I can find suggests I use parameterized queries, and I can't find any useful information on how to actually sanitize the query string itself, which seems like my only option.
For example a similar question is:
How to prevent SQL Injection in PostgreSQL JSON/JSONB field?
But I can't apply the same solution as I don't know the structure of the JSONB or the query, I can't assume the client wants to query a particular path using a particular operator, the entire JSONB query needs to be freely provided by the client.
I'm using golang, in case there are any existing libraries or code fragments that I can use.
edit: some example queries on the JSONB that the client might do:
(content->>'company') is NULL
(content->>'income')::numeric>80000
content->'company'->>'name'='EA' AND (content->>'income')::numeric>80000
content->'assets'#>'[{"kind":"car"}]'
(content->>'DOB')::TIMESTAMP<'2000-01-30T10:12:18.120Z'::TIMESTAMP
EXISTS (SELECT FROM jsonb_array_elements(content->'assets') asset WHERE (asset->>'value')::numeric > 100000)
Note that these don't cover all possible types of queries. Ideally I want any query that PostgreSQL supports on the JSONB data to be allowed. I just want to check the query to ensure it doesn't contain sql injection. For example, a simplistic and probably inadequate solution would be to not allow any ";" in the query string.
You could allow the users to specify a path within the JSON document, and then parameterize that path within a call to a function like json_extract_path_text. That is, the WHERE clause would look like:
WHERE json_extract_path_text(data, $1) = $2
The path argument is just a string, easily parameterized, which describes the keys to traverse down to the given value, e.g. 'foo.bars[0].name'. The right-hand side of the clause would be parameterized along the same rules as you're using for fixed column filtering.
I am upserting some data to a Postgres table using jOOQ's insertInto() and onDuplicateKeyUpdate() methods. I want to know later how many duplicates were in my data and hence need to return if a row was inserted or updated.
From my postgres specific research so far, I found RETURNING (not MY_TABLE.xmax = 0) AS updated to be a valid option. However, the auto-generated Java table classes from jOOQ don't seem to give me access to the system columns of postgres like xmax.
Here is my query so far:
dsl.insertInto(MY_TABLE)
.columns(
// pkey columns
MY_TABLE.SHIFT,
MY_TABLE.DATE_UTC,
MY_TABLE.TIME_UTC,
MY_TABLE.DURATION,
)
.values(
shiftId,
utcDateId,
utcTime,
duration
)
.onDuplicateKeyUpdate()
.set(MY_TABLE.DURATION, newDuration)
.returning((MY_TABLE.xmax = 0).`às`("inserted"))
.execute()
This causes the following compile time error:
Error: Kotlin: Unresolved reference: XMAX
I have rechecked my Maven jOOQ table generation configuration and I am not excluding any columns. I have also read through everything I could find on jOOQ's own website but found no useful information for this specific use-case.
Any tips on what I could do here?
In this case you should use jOOQ's SQL templating. Specifically look at the DSL.field() method. Something like this: field("my_table.xmax", int.class).eq(0).
What's the best approch to have a dynamic query like
select $dynamic_columns from table
But also prevent error like column not found and get result with available columns. considering $dynamic_columns is given by end users.
One approach would be to store the schema in java object and filter it. Again if schema is update in DB we will need to update the schema java object cache. is there any better way to handle this?
Be careful with this as it is more vulnerable to SQL injection.
Never let the user type something into a text field, instead build a
list for them to select from.
For building the list, I think the best approach is to use the JDBC method DatabaseMetaData.getColumns(...) to retrieve a list of columns for a table. I don't think there's a need to cache anything.
Can anybody help me understand the expected format of data for creating MVA (multi-value)
attributes in Sphinx?
I have a MySQL function which returns a row of comma-separated integers, collated with
GROUP_CONCAT, as a blob. I have two further MVA attributes which collate the results of a
JOIN statement, with GROUP_CONCAT, as a blob (as generated by ThinkingSphinx). These are all included in my sql_query in my sphinx.conf.
I've tried running the SQL on a small result set in the console, and it works: for all
the MVA columns, the results are a blob containing data such as:
2432,35345,342347,8975,453645
and so on. The two MVA attributes generated with the JOIN/GROUP_CONCAT combination index correctly. However, the MVA attribute generated with the MySQL function causes the
indexing to fail silently (seemingly little or no data is indexed). This is despite the query working absolutely fine in the console..
So the data format seems to be identical, but Sphinx is rejecting one of the columns. Does anybody know of any gotchas with defining MVA attributes which might help me debug
this?
I've never used thinking-sphinx (being a PHP shop here), but I don't think you should be group_concat'ing your results. From a working example in one of my sphinx.conf files:
sql_attr_multi = uint categories from query; SELECT entry_id, cat_id FROM exp_category_posts
I solved this problem eventually. It was happening because of something
which seemed unrelated: a 'sql_attr_str2ordinal' attribute which seemed to be affected
(or effect) the SQL query/indexing in ways I don't fully understand.
See: http://www.sphx.org/forum/view.html?id=2867
Fortunately, in my case I was able to remove it entirely, and indexing now seems to work.