I am use GridGain as my persistence database. I have following requirements.
Insert multiple records if key already not exists
Put multiple records if key exists or not.
For 1, I saw cache.putIfAbsent(key, value) method to insert single record if not exists. But I didn't find cache.putAllIfAbsent(Map<key, value>) like method. I can you loop to insert multiple records one by one. Is it given performance issue?
For 2, I think I can use cache.putAll(Map<key, value>) method. If it proper way?
I run server in Google cloud Kubernetes engine as thick clients.
putAll always overwrites existing records
putIfAbsent in a loop will be slower than putAll. Measure your specific use case to see by how much
If there is no requirement for ordering and atomicity, DataStreamer is a good choice. When allowOverwrite flag is false (default), you get putIfAbsent behavior, and good performance.
try (IgniteDataStreamer<Integer, String> stmr = ignite.dataStreamer("myCache")) {
stmr.allowOverwrite(false); // Don't overwrite existing data
Map<Integer, String> entries = getMydata()
stmr.addData(entries);
}
Related
Current Setup:
I have a single namespace running on the Aerospike cluster. The namespace has few sets in it.
My Use-case:
I want to copy all the records from one set (has ~100 records) to another new set (keeping the schema same) under the same namespace.
My Finding:
I did some deep dive and found out a few solutions using aql:
List down all the records from the first set and insert them one-by-one into a new set.
Pros: Simple to implement.
Cons: Time taking, and Prone to manual error.
Using asbackup/asrestore command.
Pros: It is immune to manual error.
Cons: It doesn't allow to change the name of the set during restoration, which I can't afford. Aerospike's FAQ link does provide some workaround, but again it is risky.
Help Needed:
Is there any efficient way to migrate data from one to another set, with less effort and validation? I did think of writing some java code that would scan the entire set and write those records into another set, but again that was falling under the first categories I explained earlier.
Thanks!
A record in Aerospike is stored using the hash of your key and your set name. Set name is "stored" with that record in Aerospike purely as a metadata on that record. So you can scan an entire namespace and return records belonging to that set and in the scan callback, write each of them back as new records (due to the different set name). You will have to know "your key" for each record that comes back from the scan. By default Aerospike only stores the 20byte hash digest as the key for the record. So unless you stored it explicitly in the record either with send key true or in a bin, I don't see how you would identify "your key". Storing "your key" in a bin is easiest. You may have to first update all your 100 records and add a bin that has "your key" in it. Then in scan callback, where records come in no particular order, you will be able to compose a new Key with "your key" and "new set name". You will have to write your own java code for it. (If you have "your key" in the original records - its easy to do.)
I have not tested this .. but something along these lines would work assuming original records had your key in the "mykey" bin.
client.scanAll(null, "test", "set1", new ScanCallback(){
public void scanCallback (Key key, Record record) throws AerospikeException {
String mykey = (String)(record.getValue("mykey"));
String bin1data = record.getString("bin1");
//Alternate way to get string
Key reckey = new Key("test", "set2", mykey);
client.put(null, reckey, new Bin("bin1", bin1data));
}
});
According to BigQuery docs, you can ensure data consistency providing an insertId (https://cloud.google.com/bigquery/streaming-data-into-bigquery#dataconsistency). If it's not provided, BQ will try to ensure consistency based on internals Ids and best-effort.
Using the BQ API you can do that with the row_ids param (https://google-cloud-python.readthedocs.io/en/latest/bigquery/generated/google.cloud.bigquery.client.Client.insert_rows_json.html#google.cloud.bigquery.client.Client.insert_rows_json) but I can't find the same for the Apache Beam Python SDK.
Looking into the SDK I have noticed that a 'unique_row_id' property exist, but I really don't know how to pass my param to WriteToBigQuery()
How can I write into BQ (streaming) providing a row Id for deduplication?
Update:
If you use WriteToBigQuery then it will automatically create and
insert a unique row id called insertId for you, which will be inserted to bigquery. It's handled for you, you don't need to worry about it. :)
WriteToBigQuery is a PTransform, and in it's expand method calls BigQueryWriteFn
BigQueryWriteFn is a DoFn, and in it's process method calls _flush_batch
_flush_batch is a method that then calls the BigQueryWrapper.insert_rows method
BigQueryWrspper.insert_rows creates a list of bigquery.TableDataInsertAllRequest.RowsValueListEntry objects which contain the insertId and the row data as a json object
The insertId is generated by calling the unique_row_id method which returns a value consisting of UUID4 concatenated with _ and with an auto-incremented number.
In the current 2.7.0 code, there is this happy comment; I've also verified it is true :)
https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/bigquery.py#L1182
# Prepare rows for insertion. Of special note is the row ID that we add to
# each row in order to help BigQuery avoid inserting a row multiple times.
# BigQuery will do a best-effort if unique IDs are provided. This situation
# can happen during retries on failures.
* Don't use BigQuerySink
At least, not in it's current form as it doesn't support streaming. I guess that might change.
Original (non)answer
Great question, I also looked and couldn't find a certain answer.
Apache Beam doesn't appear to use that google.cloud.bigquery client sdk you've linked to, it has some internal generated api client, but it appears to be up-to-date.
I looked at the source:
The insertall method is there https://github.com/apache/beam/blob/18d2168ee71a1b1b04976717f0f955199bb00961/sdks/python/apache_beam/io/gcp/internal/clients/bigquery/bigquery_v2_client.py#L476
I also found the insertid mentioned
https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/internal/clients/bigquery/bigquery_v2_messages.py#L1707
So if you can make an InsertAll call it will use a TableDataInsertAllRequest and pass a RowsValueListEntry
class TableDataInsertAllRequest(_messages.Message):
"""A TableDataInsertAllRequest object.
Messages:
RowsValueListEntry: A RowsValueListEntry object.
The RowsValueListEntry message is where the insertid is.
Here's the API docs for insert all
https://cloud.google.com/bigquery/docs/reference/rest/v2/tabledata/insertAll
I will look some more at this because I don't see the WriteToBigQuery() exposing this.
I suspect that the 'bigquery will remember this for at least one minute` is a pretty loose guarantee for de-duping. The docs suggest using datastore if you need transactions. Otherwise you might need to run SQL with window functions to de-dupe at runtime, or run some other de-duping jobs on bigquery.
Perhaps using batch_size parameter of WriteToBigQuery(), and running a combine (or at worst a GroupByKey) step in dataflow is a more stable way to de-dupe prior to writing.
We are going to implement gemfire for our project. We are currently syncing gemfire cache with our DB2 database. So, we are facing issue while putting DB data into cache.
To put DB data into region. I have implement com.gemstone.gemfire.cache.CacheLoader and override load method of it. As written in java doc load method will return only one Object. But for our requirement we will have to return multiple VO from load method
public List<CmDvceInvtrGemfireBean> load(LoaderHelper<CmDvceInvtrGemfireBean, CmDvceInvtrGemfireBean> helper)
throws CacheLoaderException
While returining multiple VO in form of List<CmDvceInvtrGemfireBean> gemfire region consider it's as single value.
So, when i invoke,
System.out.println("return COUNT" + cmDvceInvtrRecord.query("SELECT COUNT(*) FROM /cmDvceInvtrRecord"));
It return count of one. But i can see total 7 number of data into it.
So, I want to implement the kind of mechanism that will put all the 7 values as a separate VO in Region
Is there any way to do this using Gemfire CacheLoader?
A CacheLoader was meant to load a value only for a single entry in the GemFire Region on a cache miss. As the Javadoc states...
..creates the value for the desired key..
While a key can map to a multi-valued (e.g. an array/Collection) value, the CacheLoader can only populate a single entry.
You will have to resort to other means of populating the cache with multiple "entries" in a single operation.
Out of curiosity, why do you need (requirement?) to load multiple entries (from the DB) at once? Are you trying to minimize the number of round trips to the DB?
Also, what logic are you using to decide what VO from the DB will be loaded based on the information (i.e. key) provided in the CacheLoader?
For instance, are you somehow trying to predictably select values from the DB based on the CacheLoader key that would subsequently minimize cache misses on future Region.get(key) calls?
Sorry, I don't have a better answer for you right now, but answers to some of these questions may help me give you some ideas for alternatives.
Cheers,
John
I am developing forums on DynamoDB.
There is a table posts which contains all the posts in a thread.
I need to have a notion of sequence in the posts, i.e. I need to know which post came first and which came later.
My service would be running in a distributed env.
I am not sure if using Timestamp is the best solution for deciding the sequence, as the hosts might have slightly different times and might be off my milliseconds/ seconds.
Is there another way to do this?
Can I get DynamoDB to populate the date so it is consistent?
Or is there a sequence generator that I can use in a distributed env?
You can't use DynamoDB to auto-populate dates. You can use other services to provide you with auto-generating numbers or use DynamoDB's atomic increment to create your own UUID.
This can become a bottleneck if your forum is very successful (needs lots of numbers per second). I think you should start with timestamp and later on add complexity to your id generating (concatenate timestamp+uuid or timstamp+atomiccounter)
It is always a best practice to sync your servers clock (ntpd)
Use a dedicated sequence table. If you have only one sequence (say, PostId), then there's going to be only one row with two attributes in the table.
Yes, there's extra cost and effort of managing another table, but this is the best solution I know by far and haven't seen any one else mentioning it.
The table should have a key attribute as primary partition key, and a numeric value attribute with initial value of 1 (or whatever you want the initial value to be).
Every time you want to get the next available key, you tell DynamoDB to do this:
Increment the value where key = PostId by 1, and return the value before incrementing.
Note that this is one single atomic operation. DynamoDB handles the auto-incrementing, so there's no concurrency issues.
In code, there're more than one ways of implementing this. Here's one example:
Map<String,AttributeValue> key = new HashMap<>();
key.put("key", new AttributeValue("PostId"));
Map<String, AttributeValueUpdate> item = new HashMap<String, AttributeValueUpdate>();
item.put("value",
new AttributeValueUpdate()
.withAction(AttributeAction.ADD)
.withValue(1));
UpdateItemRequest request = new UpdateItemRequest("Sequences", key, item).withReturnValues(ReturnValue.ALL_OLD);
UpdateItemResult result = dynamoDBClient.updateItem(request);
Integer postId = Integer.parseInt(result.getAttributes().get("value").getN()); // <- this is the sequential ID you want to set to your post
Another variation of Chen's suggestion is to have strict ordering of posts within a given Forum Thread, as opposed to globally across all Threads. One way to do this is to have a Reply table with the Hash key of ThreadId, and a range key of ReplyId. The ReplyId would be a Number type attribute starting at 0. Every time someone replies, your app does a Query on the Reply table for the one most recent reply on that thread (ScanIndexForward: false, Limit: 1, ThreadId: ). To insert your new reply use the ReplyId of the one returned in the Query, + 1. Then use PutItem, using a Conditional Write, so that if someone else replies at the same time, an error will be returned, and your app can start again with the query.
If you want the simplest initial solution possible, then the timestamp+uuid concatenation Chen suggests is the simplest approach. A global atomic counter item will be a scaling bottleneck, as Chen mentions, and based on what you've described, a global sequence number isn't required for your app.
I am looking for the fastest way to check for the existence of an object.
The scenario is pretty simple, assume a directory tool, which reads the current hard drive. When a directory is found, it should be either created, or, if already present, updated.
First lets only focus on the creation part:
public static DatabaseDirectory Get(DirectoryInfo dI)
{
var result = DatabaseController.Session
.CreateCriteria(typeof (DatabaseDirectory))
.Add(Restrictions.Eq("FullName", dI.FullName))
.List<DatabaseDirectory>().FirstOrDefault();
if (result == null)
{
result = new DatabaseDirectory
{
CreationTime = dI.CreationTime,
Existing = dI.Exists,
Extension = dI.Extension,
FullName = dI.FullName,
LastAccessTime = dI.LastAccessTime,
LastWriteTime = dI.LastWriteTime,
Name = dI.Name
};
}
return result;
}
Is this the way to go regarding:
Speed
Separation of Concern
What comes to mind is the following: A scan will always be performed "as a whole". Meaning, during a scan of drive C, I know that nothing new gets added to the database (from some other process). So it MAY be a good idea to "cache" all existing directories prior to the scan, and look them up this way. On the other hand, this may be not suitable for large sets of data, like files (which will be 600.000 or more)...
Perhaps some performance gain can be achieved using "index columns" or something like this, but I am not so familiar with this topic. If anybody has some references, just point me in the right direction...
Thanks,
Chris
PS: I am using NHibernate, Fluent Interface, Automapping and SQL Express (could switch to full SQL)
Note:
In the given problem, the path is not the ID in the database. The ID is an auto-increment, and I can't change this requirement (other reasons). So the real question is, what is the fastest way to "check for the existance of an object, where the ID is not known, just a property of that object"
And batching might be possible, by selecting a big group with something like "starts with C:Testfiles\" but the problem then remains, how do I know in advance how big this set will be. I cant select "max 1000" and check in this buffered dictionary, because i might "hit next to the searched dir"... I hope this problem is clear. The most important part, is, is buffering really affecting performance this much. If so, does it make sense to load the whole DB in a dictionary, containing only PATH and ID (which will be OK, even if there are 1.000.000 object, I think..)
First off, I highly recommend that you (anyone using NH, really) read Ayende's article about the differences between Get, Load, and query.
In your case, since you need to check for existence, I would use .Get(id) instead of a query for selecting a single object.
However, I wonder if you might improve performance by utilizing some knowledge of your problem domain. If you're going to scan the whole drive and check each directory for existence in the database, you might get better performance by doing bulk operations. Perhaps create a DTO object that only contains the PK of your DatabaseDirectory object to further minimize data transfer/processing. Something like:
Dictionary<string, DirectoryInfo> directories;
session.CreateQuery("select new DatabaseDirectoryDTO(dd.FullName) from DatabaseDirectory dd where dd.FullName in (:ids)")
.SetParameterList("ids", directories.Keys)
.List();
Then just remove those elements that match the returned ID values to get the directories that don't exist. You might have to break the process into smaller batches depending on how large your input set is (for the files, almost certainly).
As far as separation of concerns, just keep the operation at a repository level. Have a method like SyncDirectories that takes a collection (maybe a Dictionary if you follow something like the above) that handles the process for updating the database. That way your higher application logic doesn't have to worry about how it all works and won't be affected should you find an even faster way to do it in the future.