Hello,
I continuously stream massive amounts of market data from ClickHouse and after profiling my app I discovered that most ticks are spent in stream parsing, i..e:

       const rows = await this.#client.query({
            query: `
                SELECT
                *
            FROM...
   `})
  
       const stream = rows.stream()              

        stream.on('data', async (rows) => {
            for (const row of rows) {
                const [
                    event,
                    time,
                    ...
                ] = await row.json()
      })

I briefly checked the module source code, and looks like quite expensive JSON.parse() is used internally even when we call text().

My question would be **is there a faster way to parse incoming stream maybe by using some raw output format? ** Or am I using reading streams wrong?