Possible new Search API v1.1 Errata


#1

Hello, I think there may be some kind of error in the string returned as ‘next_results’ when executing a search with the ‘result_type’ set as ‘mixed’.

The ‘max_id’ on the ‘next_results’ returned is set to the Id of the oldest tweet received, but it consider the tweets with ‘result_type’ as ‘popular’, and it should only consider the ones with ‘result_type’ as ‘recent’ because otherwise the oldest tweet would be probably one of the popular ones and when we use the ‘next_results’ string for our next query, the tweets returned will be much older than expected.

For example for the query:
https://api.twitter.com/1.1/search/tweets.json?q=twitter&result_type=mixed&count=1

The result obtained would be:

{
	statuses: [4]
		0:  {
			metadata: {
				result_type: "popular"
				iso_language_code: "en"
			}-
			created_at: "Wed May 08 13:24:12 +0000 2013"
			id: 332123772679249900
			id_str: "332123772679249920"
			...
			}
		1:  {
			metadata: {
				result_type: "popular"
				iso_language_code: "nl"
			}-
			created_at: "Wed May 08 19:00:04 +0000 2013"
			id: 332208295903391740
			id_str: "332208295903391744"
			...
		}-
		2:  {
			metadata: {
				result_type: "popular"
				iso_language_code: "en"
			}-
			created_at: "Wed May 08 07:38:46 +0000 2013"
			id: 332036840750207000
			id_str: "332036840750206976"
			...
		}-
		3:  {
			metadata: {
				result_type: "recent"
				iso_language_code: "en"
			}-
			created_at: "Thu May 09 11:04:40 +0000 2013"
			id: 332451045080965100
			id_str: "332451045080965121"
			...
		}-
		-
	search_metadata: {
		max_id: 332451045080965100
		max_id_str: "332451045080965121"
		next_results: "?max_id=332036840750206975&q=twitter&count=1&include_entities=1&result_type=mixed"
		query: "twitter"
		refresh_url: "?since_id=332451045080965121&q=twitter&result_type=mixed&include_entities=1"
		count: 1
		since_id: 0
		since_id_str: "0"
	}-
}

As you can see [Line 47] the ‘next_results’ have as ‘max_id’ the id corresponding with the third popular tweet [Line 30] not with the last one of the ‘recent’ ones [Line 40]. So when we execute the next query the tweets that we get will be tweets older than the third popular and we will be probably missing many tweets.

I don’t know if this is the desired behavior but it seems very odd to me.
Best regards.


#2

Is this problem resolved?