cardano-graphql
cardano-graphql copied to clipboard
input-output-hk
searching for transactions by address with a very large number of transactions is slow
this is graphql
query getAddressTransactions($address: String!) {
transactions(
limit: 100
where: {
_or:[
{ inputs: { address: { _eq: $address } } }
{ outputs: { address: { _eq: $address } } }
]
}
order_by:{
includedAt:desc
}
) {
block {
number
}
hash
fee
includedAt
inputs {
address
value
}
outputs {
address
value
}
}
}
{
"address":"DdzFFzCqrhstmqBkaU98vdHu6PdqjqotmgudToWYEeRmQKDrn4cAgGv9EZKtu1DevLrMA1pdVazufUCK4zhFkUcQZ5Gm88mVHnrwmXvT"
}
It takes more than 10 seconds to return results, Is there any other way to get the transactions of an address?
It's proportional to the number of transactions, and that address has 75483. We're aware of the issue, and will be looking to improve the performance in this area.
We're aware of the issue, and will be looking to improve the performance in this area.
Are there any updates to this? By the way, if I want to get latest 10 transactions of the account using order_by: { block:{number:desc}} along with the address summary it also increases the query time from 5 seconds up to 30 seconds even if the amount of transactions is too low (approx. 10-30)
Query example
query getAddressSummary($address: String!) {
caTxNum: transactions_aggregate(
where: {
_or: [
{ inputs: { address: { _eq: $address } } }
{ outputs: { address: { _eq: $address } } }
]
}
) {
aggregate {
sum{fee}
count# caTxNum
}
}
caTotalInput: transactions_aggregate(
where: { inputs: { address: { _eq: $address } } }
) {
aggregate {
sum {
totalOutput # caTotalInput
}
}
}
caTotalOutput: transactions_aggregate(
where: { outputs: { address: { _eq: $address } } }
) {
aggregate {
sum {
totalOutput # caTotalOutput
}
}
}
caBalance: utxos_aggregate(where: { address: { _eq: $address } }) {
aggregate {
sum {
value # caBalance
}
}
}
caTxList: transactions(limit: 10
order_by: { block:{number:desc}}
where: {
_or: [
{ inputs: { address: { _eq: $address } } }
{ outputs: { address: { _eq: $address } } }
]
}
) {
hash # ctbId
includedAt # ctbTimeIssued
fee # ctbFees
inputs {
# ctbInputs
address # ctaAddress
value # ctaAmount
sourceTxHash # ctaTxHash
sourceTxIndex # ctaTxIndex
}
outputs {
address # ctaAddress
value # ctaAmount
txHash # ctaTxHash
index
}
inputs_aggregate {
aggregate {
sum {
value # ctbInputSum
}
}
}
outputs_aggregate {
aggregate {
sum {
value # ctbOutputSum
}
}
}
}
}
That makes sense as it still has to get all the transactions first so it can order them. Even if the DB has an index applied.
A tip, make sure the device you're using has high IOPS. E.g, I tried using AWS and the IOPS was so slow that most queries timed out. The price to reserve enough IOPS for it to be usable wasn't worth it.
We're aware of the issue, and will be looking to improve the performance in this area.
Are there any updates to this? By the way, if I want to get latest 10 transactions of the account using
order_by: { block:{number:desc}}along with the address summary it also increases the query time from 5 seconds up to 30 seconds even if the amount of transactions is too low (approx. 10-30)Query example
query getAddressSummary($address: String!) { caTxNum: transactions_aggregate( where: { _or: [ { inputs: { address: { _eq: $address } } } { outputs: { address: { _eq: $address } } } ] } ) { aggregate { sum{fee} count# caTxNum } } caTotalInput: transactions_aggregate( where: { inputs: { address: { _eq: $address } } } ) { aggregate { sum { totalOutput # caTotalInput } } } caTotalOutput: transactions_aggregate( where: { outputs: { address: { _eq: $address } } } ) { aggregate { sum { totalOutput # caTotalOutput } } } caBalance: utxos_aggregate(where: { address: { _eq: $address } }) { aggregate { sum { value # caBalance } } } caTxList: transactions(limit: 10 order_by: { block:{number:desc}} where: { _or: [ { inputs: { address: { _eq: $address } } } { outputs: { address: { _eq: $address } } } ] } ) { hash # ctbId includedAt # ctbTimeIssued fee # ctbFees inputs { # ctbInputs address # ctaAddress value # ctaAmount sourceTxHash # ctaTxHash sourceTxIndex # ctaTxIndex } outputs { address # ctaAddress value # ctaAmount txHash # ctaTxHash index } inputs_aggregate { aggregate { sum { value # ctbInputSum } } } outputs_aggregate { aggregate { sum { value # ctbOutputSum } } } } }
Also, that query is very complex. I'd split it up into seperate statements and call them separately. It's easy enough for it to get choked if you ask it to do to much at once.
In my case there is no a lot of transactions on my addresses (around 1000), but querying for utxos_aggregate and transactions takes a lot of time
In pg node exporter I can see that queries return millions of rows. Seems it's not optimal to return a lot of data from postgres and then process it in hasura
hi, we nowadays get 10 sec queries even for addresses with 0 transactions
'{"query": "{ transactions ( limit: 25 where: { _or: [ { inputs: { address: { _eq: \"addr1q88teuwqrs04n6h68l5t0s54tgn2r0yr7hu5cslpzg75ejxmctnagnrgtk7kwwhn0x6xkrnk9sh2w6lqkf38etlqrzzqaf4qdf\" } } } { outputs: { address: { _eq: \"addr1q88teuwqrs04n6h68l5t0s54tgn2r0yr7hu5cslpzg75ejxmctnagnrgtk7kwwhn0x6xkrnk9sh2w6lqkf38etlqrzzqaf4qdf\" } } } ] }) { block { hash number } } }" }'
