Related
Let's assume that we have a file with the values as seen bellow:
% head test.csv
20220601,A,B,1
20220530,A,B,1
And we want to add two new columns, one with the date minus 1 day and one with minus 7 days, resulting the following:
% head new_test.csv
20220601,A,B,20220525,20220531,1
20220530,A,B,20220523,20220529,1
The awk that was used to produce the above is:
% awk 'BEGIN{FS=OFS=","} { a="date -d \"$(date -d \""$1"\") -7 days\" +'%Y%m%d'"; a | getline st ; close(a) ;b="date -d \"$(date -d \""$1"\") -1 days\" +'%Y%m%d'"; b | getline cb ; close(b) ;print $1","$2","$3","st","cb","$4}' test.csv > new_test.csv
But after applying the above in a large file with more than 100K lines it runs for 20 minutes, is there any way to optimize the awk?
One GNU awk approach:
awk '
BEGIN { FS=OFS=","
secs_in_day = 60 * 60 * 24
}
{ dt = mktime( substr($1,1,4) " " substr($1,5,2) " " substr($1,7,2) " 12 0 0" )
dt1 = strftime("%Y%m%d",dt - secs_in_day )
dt7 = strftime("%Y%m%d",dt - (secs_in_day * 7) )
print $1,$2,$3,dt7,dt1,$4
}
' test.csv
This generates:
20220601,A,B,20220525,20220531,1
20220530,A,B,20220523,20220529,1
NOTES:
requires GNU awk for the mktime() and strftime() functions; see GNU awk time functions for more details
other flavors of awk may have similar functions, ymmv
You can try using function calls, it is faster than calling the .
awk -F, '
function cmd1(date){
a="date -d \"$(date -d \""date"\") -1days\" +'%Y%m%d'"
a | getline st
return st
close(a)
}
function cmd2(date){
b="date -d \"$(date -d \""date"\") -7days\" +'%Y%m%d'"
b | getline cm
return cm
close(b)
}
{
$5=cmd1($1)
$6=cmd2($1)
print $1","$2","$3","$5","$6","$4
}' OFS=, test > newFileTest
I executed this against a file with 20000 records in seconds, compared to the original awk which took around 5 minutes.
Need to find the difference between TradeCloseTime and TradeOpenTime time in dd:hh:mm format for the Exposure column in the following script.
Also the script runs super slow (~4 mins for 800 rows of json, on Core i7 16gb RAM machine)
#!/bin/bash
echo "TradeNo, TradeOpenType, TradeCloseType, TradeOpenSource, TradeCloseSource, TradeOpenTime, TradeCloseTime, PNL, Exposure" > tradelist.csv
tradecount=$(jq -r '.performance.numberOfTrades|tonumber' D.json)
for ((i=0; i<$tradecount; i++))
do
tradeNo=$(jq -r '.trades['$i']|[.tradeNo][]|tonumber' D.json)
entrySide=$(jq -r '.trades['$i'].orders[0]|[.side][]' D.json)
exitSide=$(jq -r '.trades['$i'].orders[1]|[.side][]' D.json)
entrySource=$(jq -r '.trades['$i'].orders[0]|[.source][]' D.json)
exitSource=$(jq -r '.trades['$i'].orders[1]|[.source][]' D.json)
tradeEntryTime=$(jq -r '.trades['$i'].orders[0]|[.placedTime][]' D.json | tr -d 'Z' | tr -s 'T' ' ')
tradeExitTime=$(jq -r '.trades['$i'].orders[1]|[.placedTime][]' D.json | tr -d 'Z' | tr -s 'T' ' ')
profitPercentage=$(jq -r '(.trades['$i']|[.profitPercentage][0]|tonumber)*(100)' D.json)
echo $tradeNo","$entrySide","$exitSide","$entrySource","$exitSource","$tradeEntryTime","$tradeExitTime","$profitPercentage | tr -d '"' >> tradelist.csv
done
json file looks like this
{"market":{"exchange":"BINANCE_FUTURES","coinPair":"BTC_USDT"},"strategy":{"name":"","type":"BACKTEST","candleSize":15,"lookbackDays":6,"leverageLong":1.00000000,"leverageShort":1.00000000,"strategyName":"ABC","strategyVersion":35,"runNo":"002","source":"Personal"},"strategyParameters":[{"name":"DurationInput","value":"87.0"}],"openPositionStrategy":{"actionTime":"CANDLE_CLOSE","maxPerSignal":1.00000000},"closePositionStrategy":{"actionTime":"CANDLE_CLOSE","minProfit":"NaN","stopLossValue":0.07000000,"stopLossTrailing":true,"takeProfit":0.01290000,"takeProfitDeviation":"NaN"},"performance":{"startTime":"2019-01-01T00:00:00Z","endTime":"2021-11-24T00:00:00Z","startAllocation":1000.00000000,"endAllocation":3478.58904150,"absoluteProfit":2478.58904150,"profitPerc":2.47858904,"buyHoldRatio":0.62426630,"buyHoldReturn":4.57228387,"numberOfTrades":744,"profitableTrades":0.67833109,"maxDrawdown":-0.20924885,"avgMonthlyProfit":0.05242718,"profitableMonths":0.70370370,"avgWinMonth":0.09889897,"avgLoseMonth":-0.05275563,"startPrice":null,"endPrice":57623.08000000},"trades":[{"tradeNo":0,"profit":-5.48836165,"profitPercentage":-0.00549085,"accumulatedBalance":994.51163835,"compoundProfitPerc":-0.00548836,"orders":[{"side":"Long","placedTime":"2019-09-16T21:15:00Z","placedAmount":0.09700000,"filledTime":"2019-09-16T21:15:00Z","filledAmount":0.09700000,"filledPrice":10300.49000000,"commissionPaid":0.39965901,"source":"SIGNAL"},{"side":"CloseLong","placedTime":"2019-09-17T19:15:00Z","placedAmount":0.09700000,"filledTime":"2019-09-17T19:15:00Z","filledAmount":0.09700000,"filledPrice":10252.13000000,"commissionPaid":0.39778264,"source":"SIGNAL"}]},{"tradeNo":1,"profit":-3.52735800,"profitPercentage":-0.00356403,"accumulatedBalance":990.98428035,"compoundProfitPerc":-0.00901572,"orders":[{"side":"Long","placedTime":"2019-09-19T06:00:00Z","placedAmount":0.10000000,"filledTime":"2019-09-19T06:00:00Z","filledAmount":0.10000000,"filledPrice":9893.16000000,"commissionPaid":0.39572640,"source":"SIGNAL"},{"side":"CloseLong","placedTime":"2019-09-19T06:15:00Z","placedAmount":0.10000000,"filledTime":"2019-09-19T06:15:00Z","filledAmount":0.10000000,"filledPrice":9865.79000000,"commissionPaid":0.39463160,"source":"SIGNAL"}]},{"tradeNo":2,"profit":-5.04965308,"profitPercentage":-0.00511770,"accumulatedBalance":985.93462727,"compoundProfitPerc":-0.01406537,"orders":[{"side":"Long","placedTime":"2019-09-25T10:15:00Z","placedAmount":0.11700000,"filledTime":"2019-09-25T10:15:00Z","filledAmount":0.11700000,"filledPrice":8430.00000000,"commissionPaid":0.39452400,"source":"SIGNAL"},{"side":"CloseLong","placedTime":"2019-09-25T10:30:00Z","placedAmount":0.11700000,"filledTime":"2019-09-25T10:30:00Z","filledAmount":0.11700000,"filledPrice":8393.57000000,"commissionPaid":0.39281908,"source":"SIGNAL"}]}
You can do it all (extracts, conversions and formatting) with one jq call:
#!/bin/sh
echo 'TradeNo,TradeOpenType,TradeCloseType,TradeOpenSource,TradeCloseSource,TradeOpenTime,TradeCloseTime,PNL,Exposure'
query='
.trades[]
| [
.tradeNo,
.orders[0].side,
.orders[1].side,
.orders[0].source,
.orders[1].source,
(.orders[0].placedTime | fromdate | strftime("%Y-%m-%d %H:%M:%S")),
(.orders[1].placedTime | fromdate | strftime("%Y-%m-%d %H:%M:%S")),
.profitPercentage * 100,
(
(.orders[1].placedTime | fromdate) - (.orders[0].placedTime | fromdate)
| (. / 86400 | floor | tostring) + (. % 86400 | strftime(":%H:%M"))
)
]
|#csv
'
jq -r "$query" < D.json > tradelist.csv
example of JSON (cleaned of all irrelevant keys):
{
"trades": [
{
"tradeNo": 0,
"profitPercentage": -0.00549085,
"orders": [
{
"side": "Long",
"placedTime": "2018-12-16T21:34:46Z",
"source": "SIGNAL"
},
{
"side": "CloseLong",
"placedTime": "2019-09-17T19:15:00Z",
"source": "SIGNAL"
}
]
}
]
}
output:
TradeNo,TradeOpenType,TradeCloseType,TradeOpenSource,TradeCloseSource,TradeOpenTime,TradeCloseTime,PNL,Exposure
0,"Long","CloseLong","SIGNAL","SIGNAL","2018-12-16 21:34:46","2019-09-17 20:15:00",-0.549085,"274:22:40"
If you want to get rid of the double quotes that jq adds when generating a CSV (which are completely valid, but you need a real parser to read the CSV) then you can replace #csv with #tsv and post-process the output with tr '\t' ',', like this:
query='
...
|#tsv
'
jq -r "$query" < D.json | tr '\t' ',' > tradelist.csv
and you'll get:
TradeNo,TradeOpenType,TradeCloseType,TradeOpenSource,TradeCloseSource,TradeOpenTime,TradeCloseTime,PNL,Exposure
0,Long,CloseLong,SIGNAL,SIGNAL,2018-12-16 21:34:46,2019-09-17 20:15:00,-0.549085,274:22:40
note: This method of getting rid of the " in the CSV is only accurate when there is no \n \t \r \ , or " characters in the input data.
Regarding the main question (regarding computing time differences), you're in luck as jq provides the built-in function fromdateiso8601 for converting ISO times to "the
number of seconds since the Unix epoch (1970-01-01T00:00:00Z)".
With your JSON sample,
.trades[]
| [ .orders[1].placedTime, .orders[0].placedTime]
| map(fromdateiso8601)
| .[0] - .[1]
produces the three differences:
79200
900
900
And here's a function for converting seconds to "hh:mm:ss" format:
def hhmmss:
def l: tostring | if length < 2 then "0\(.)" else . end;
(. % 60) as $ss
| ((. / 60) | floor) as $mm
| (($mm / 60) | floor) as $hh
| ($mm % 60) as $mm
| [$hh, $mm, $ss] | map(l) | join(":");
I prefer using an intermediate structure of the "entry" and "exit" JSON. This helps with debugging the jq commands. Formatted for readability over performance:
#!/usr/bin/env bash
echo "TradeNo,TradeOpenType,TradeCloseType,TradeOpenSource,TradeCloseSource,TradeOpenTime,TradeCloseTime,PNL,Exposure" > tradelist.csv
jq -r '
.trades[]
|{tradeNo,
profitPercentage,
entry:.orders[0],
exit:.orders[1],
entryTS:.orders[0].placedTime|fromdate,
exitTS:.orders[1].placedTime|fromdate}
|[.tradeNo,
.entry.side,
.exit.side,
.entry.source,
.exit.source,
(.entry.placedTime|strptime("%Y-%m-%dT%H:%M:%SZ")|strftime("%Y-%m-%d %H:%M:%S")),
(.exit.placedTime|strptime("%Y-%m-%dT%H:%M:%SZ")|strftime("%Y-%m-%d %H:%M:%S")),
(.profitPercentage*100),
(.exitTS-.entryTS|todate|strptime("%Y-%m-%dT%H:%M:%SZ")|strftime("%d:%H:%M"))]|#csv
' D.json | tr -d '"' >> tradelist.csv
WARNING: This formatting assumes Exposure is LESS THAN 1 MONTH. Good luck with that!
I have the following ouput
[
"notimportant",
[
"val1",
"val2",
...,
"valn"
]
]
I'm trying to store every value into a bash string, using jq I tried this
out=''
req=$(curl -s $url)
len=$(echo $req | jq length )
for (( i = 0; i < $len; i++ )); do
element=$(echo $req | jq '.[1]' | jq --argjson i "$i" '.[$i]')
out=${element}\n${out}
done
which feels clunky and also has a slow performance. I'm trying to dump the values at once without looping on all the elements
With an array:
mapfile -t arr < <(curl -s "$url" | jq -r '.[1] | .[]')
declare -p arr
Do you want the values separate by TAB or NEWLINE characters in a single variable? The #tsv function is useful for controlling output:
outTABS=$(curl -s "$url" | jq -r '.[1]|.|#tsv')
outLINE=$(curl -s "$url" | jq -r '.[1]|.[]|[.]|#tsv')
> echo "$outTABS"
val1 val2 valn
> echo "$outLINE"
val1
val2
valn
I have data in json format that logs timestamps (hh:mm in 24h format) with an event (In/Out). My goal is to add up all the time differences between an "IN" event and the next "OUT" event.
For simplification I assume that there are no inconsistencies (The first element is always an "IN" and each "IN" is followed by an "OUT"). Exception: If the last element is an "IN", the calculation has to be done between the current time and the timestamp from the last "IN" event.
This is my script so far, which calculates all the timespans, also between an OUT and an IN event. But I need only those that are inbetween an IN and OUT event.
Any tips what might be more useful here are welcome !
#!/bin/bash
JSON='{ "times": [ [ "7:43", "IN" ], [ "8:26", "OUT" ], [ "8:27", "IN" ], [ "9:12", "OUT" ], [ "9:14", "IN" ], [ "9:22", "OUT" ], [ "9:23", "IN " ], [ "12:12", "OUT" ], [ "13:12", "IN" ] ]}'
IN_TIMES=$(jq '.times | to_entries | .[] | select(.value[1]| tostring | contains("IN")) | .value[0]' <<< "$JSON")
OUT_TIMES=$(jq '.times | to_entries | .[] | select(.value[1]| tostring | contains("OUT")) | .value[0]' <<< "$JSON")
ALL_TIMES=$(jq -r '.times| to_entries | .[] | .value[0]' <<< "$JSON")
prevtime=0
count=0
for i in $(echo $ALL_TIMES | sed "s/ / /g")
do
if [[ "$count" -eq 0 ]]; then
(( count++ ))
prevtime=$i
continue
else
(( count++ ))
fi
time1=`date +%s -d ${prevtime}`
time2=`date +%s -d ${i}`
diffsec=`expr ${time2} - ${time1}`
echo From $prevtime to $i: `date +%H:%M -ud #${diffsec}`
prevtime=$i
done
Here's an only-jq solution that only calls jq once.
Please note, though, that it may need tweaking to take into account time zone considerations, error-handling, and potentially other complications:
def mins: split(":") | map(tonumber) | .[0] * 60 + .[1];
def diff: (.[1] - .[0]) | if . >= 0 then . else 24*60 + . end;
def now_mins: now | gmtime | .[3] * 60 + .[4];
def pairs:
range(0; length; 2) as $i | [.[$i], .[$i+1] ];
def sigma(s): reduce s as $s (0; . + $s);
.times
| map( .[0] |= mins )
| if .[-1][1] == "IN" then . + [ [now_mins, "OUT"] ] else . end
| sigma(pairs | map(.[0]) | diff)
Since you measure times up to the minute, it is enough to compute minutes without messing up with the command date. I have an awk solution:
awk -F: -vIRS=" " -vfmt="From %5s to %5s: %4u minutes\n" \
'{this=$1*60+$2}a{printf(fmt,at,$0,this-a);a=0;next}{a=this;at=$0}\
END{if(a){$0=strftime("%H:%M");printf(fmt,at,$0,$1*60+$2-a)}}' <<<"$ALL_TIMES"
which works by defining a colon as field separator and a space as record separator. In this way we get a separate record with two fields for each time. Then
{this=$1*60+$2} : We compute how many minutes there are in the current record and put them in the variable this.
a{printf(fmt,at,$0,this-a);a=0;next} : If the (initially empty) variable a is not null nor zero, we are reading an OUT entry, so we print what we want, set a to zero because the next field will be an IN entry, and we continue to the next record.
{a=this;at=$0} : Otherwise, we are reading an IN entry, and set a to its minutes and at to its string representation (needed we will print it, as per previous case).
END{if(a){$0=strftime("%H:%M");printf(fmt,at,$0,$1*60+$2-a)}} : at the end, if we still have some dangling IN data, we set $0 to be the properly formatted current time and print what we want.
All done.
With Xidel and a little XQuery magic this is rather simple:
#!/bin/bash
JSON='{"times": [["7:43", "IN"], ["8:26", "OUT"], ["8:27", "IN"], ["9:12", "OUT"], ["9:14", "IN"], ["9:22", "OUT"], ["9:23", "IN "], ["12:12", "OUT"], ["13:12", "IN"]]}'
xidel -s - --xquery '
let $in:=$json/(times)()[contains(.,"IN")](1) ! time(
substring(
"00:00:00",
1,
8-string-length(.)
)||.
),
$out:=$json/(times)()[contains(.,"OUT")](1) ! time(
substring(
"00:00:00",
1,
8-string-length(.)
)||.
)
for $x at $i in $out return
concat(
"From ",
$in[$i],
" to ",
$x,
": ",
$x - $in[$i] + time("00:00:00")
)
' <<< "$JSON"
$in:
00:07:43
00:08:27
00:09:14
00:09:23
00:13:12
$out:
00:08:26
00:09:12
00:09:22
00:12:12
Output:
From 00:07:43 to 00:08:26: 00:00:43
From 00:08:27 to 00:09:12: 00:00:45
From 00:09:14 to 00:09:22: 00:00:08
From 00:09:23 to 00:12:12: 00:02:49
I have data in this format in a file:
{"field1":249449,"field2":116895,"field3":1,"field4":"apple","field5":42,"field6":"2019-07-01T00:00:10","metadata":"","frontend":""}
{"field1":249448,"field2":116895,"field3":1,"field4":"apple","field5":42,"field6":"2019-07-01T00:00:10","metadata":"","frontend":""}
{"field1":249447,"field2":116895,"field3":1,"field4":"apple","field5":42,"field6":"2019-07-01T00:00:10","metadata":"","frontend":""}
{"field1":249443,"field2":116895,"field3":1,"field4":"apple","field5":42,"field6":"2019-07-01T00:00:10","metadata":"","frontend":""}
{"field1":249449,"field2":116895,"field3":1,"field4":"apple","field5":42,"field6":"2019-07-01T00:00:10","metadata":"","frontend":""}
Here, each entry represents a row. I want to have a count of the rows with respect to the value in field one, like:
249449 : 2
249448 : 1
249447 : 1
249443 : 1
How can I get that?
with awk
$ awk -F'[,:]' -v OFS=' : ' '{a[$2]++} END{for(k in a) print k, a[k]}' file
You can use the jq command line tool to interpret JSON data. uniq -c counts the number of occurences.
% jq .field1 < $INPUTFILE | sort | uniq -c
1 249443
1 249447
1 249448
2 249449
(tested with jq 1.5-1-a5b5cbe on linux xubuntu 18.04 with zsh)
Here's an efficient jq-only solution:
reduce inputs.field1 as $x ({}; .[$x|tostring] += 1)
| to_entries[]
| "\(.key) : \(.value)"
Invocation: jq -nrf program.jq input.json
(Note in particular the -n option.)
Of course if an object-representation of the counts is satisfactory, then
one could simply write:
jq -n 'reduce inputs.field1 as $x ({}; .[$x|tostring] += 1)' input.json
Using datamash and some shell utils, change the non-data delimiters to squeezed tabs, count field 3, (it'd be field 2, but there's a leading tab), reverse, then pretty print as per OP spec:
tr -s '{":,}' '\t' < file | datamash -sg 3 count 3 | tac | xargs printf '%s : %s\n'
Output:
249449 : 2
249448 : 1
249447 : 1
249443 : 1